v

SYPHILIS IN THE IETIOLOGY ANDDIAGNOSIS OF TUBERCULOSIS *

By S. ROODHOUSE GLOYNE, M.D., D.P.H., Pathologist, City ofLondon Hospital for Diseases of the Heart and Lungs, VictoriaPark.

THE confusion between syphilis and tuberculosis musthave been a sore trial to the early morbid anatomists, andeven to the clinicians. Some of the older writers believedthat syphilis tended to protect a patient against tubercu-losis, others that it predisposed him to infection. Themore precise methods of laboratory diagnosis have, it istrue, produced in recent years a wealth of new material,but they have by no means settled for us this vexed ques-tion. Rather they appear to have complicated certainaspects of it, as, for example, the interpretation of theWassermann reaction in tuberculosis. Before attemptingto draw any conclusions, therefore, it is essential tocollect and examine the data.

LETIOLOGY

Beginning with the question of aetiology, two mainissues are involved in the inquiry: (I) the possibilityof syphilis preparing the soil for the seed of tuberculosisand (2) the effect of intercurrent syphilis upon an activeor quiescent tuberculosis. Except, however, in the caseof tuberculous patients who actually develop a chancreand signs of secondary syphilis during the period ofobservation, and in the case of childhood tuberculosiswith a syphilitic heredity, it is often impossible to decidewhich is the antecedent disease. It will be thereforebest to record first those cases which show evidence ofboth diseases irrespective of priority. Let us take firstthe record of positive Wassermann reactions in tubercu-losis institutions and in tuberculous patients generally.

Hollander and F. C. Narr 1 have analysed the pub-lished records of different observers. These records show

* Based on a paper read at General Meeting June ist, 1928.293 Y 2

copyright. on F

ebruary 22, 2022 by guest. Protected by

http://sti.bmj.com

/B

r J Vener D

is: first published as 10.1136/sti.4.4.293 on 1 October 1928. D

ownloaded from

BRITISH JOURNAL OF VENEREAL DISEASES

a definitely positive Wassermann reaction in IO-36 percent. of 6,324 tuberculous persons. The highest recordin the series is that of Letulle and his colleagues, whiclhdeals with French patients. Their proportion of positivesis I9 per cent. The remainder of the records in the -eriesis American, and the figures range from 2 to I7 per cent.of positives. It would be easy to add still more recordsto this list. Two, however, will suffice. The first is thatof Adelung,2 of America, which I quote because all hispatients were pulmonary cases with tubercle bacilli in thesputum. He found 8-7 per cent. of positive Wassermannreactions in I95 patients. The second, quoted becauseit is the only one I can find concerning Great Britain, isthat of Bowman.3 Examining 500 patients in the Glas-gow City Sanatorium, he found 9-42 per cent. of positivesin 223 pulmonary cases and 5o05 per cent. in 277 non-pulmonary cases. As a rough control of all these figuresmay be quoted Kilduffe's series 4 of I2 per cent. ofpositives in 484 unselected patients attending a generalhospital in Pittsburg, where presumably patients ofdifferent European and American stocks are represented.This figure is slightly higher than that given (8-IO percent.) by the Royal Commission on Venereal Diseases inthis country in I9I6.

Obviously, it is not possible to make a very strict com-parison between these figures, which are compiled fromdifferent types of clinical materials and from differentsources. For instance, many of them are crude figures,including acquired and hereditary syphilis. Those inter-ested in this subject are referred to a very thorough paperby Ritter,5 in which a large collection of data from morethan fifty sources is given.At Victoria Park it is not our practice to make a

Wassermann test on all patients. On looking up therecords I find that tests have been carried out on I37patients from the tuberculosis and observation wardsduring the last four years. Of these twenty-six (I9 percent.) were positive, but only six of them (4 per cent. ofthe total) showed definite and conclusive evidence oftuberculosis, the others being probably, if not certainly,non-tuberculous. Of these six one was a fatal case ofpulmonary tuberculosis and another a fatal case oftuberculous peritonitis in a child; the other four weremales with the chronic middle-aged type of pulmonary

294

copyright. on F

ebruary 22, 2022 by guest. Protected by

http://sti.bmj.com

/B

r J Vener D

is: first published as 10.1136/sti.4.4.293 on 1 October 1928. D

ownloaded from

DIAGNOSIS OF TUBERCULOSIS

tuberculosis, one having a tuberculous larynx andanother tabes dorsalis. Obviously, some selection was atwork in having Wassermann tests done on these cases,but, whatever their interpretation, they do not emphasisesyphilis as a predisposing cause of tuberculosis. Tosum up, the figures from various sources quoted aboveshow that the percentage of positive Wassermannreactions in institutions dealing with cases of tuberculosisand with tuberculous patients generally is, broadly speak-ing, little higher than that found in the general population.The next question which arises is, Do phthisis cases

with a syphilitic history exhibit a form of tuberculosisdiffering from that usually found in non-specific patients KSome writers 6 have described a caseous pneumonic typeof tuberculosis and a liability to early lardaceous diseaseSergent, a fibroid form of pulmonary tuberculosis ofslow growth and associated with emphysema and arterio-sclerosis; Morton,8 again, an increased liability to laryn-geal tuberculosis. My experience at Victoria Park leadsme to agree in general with the last two-named views, butnot with the first. I have not seen the caseous pneumonictype associated with syphilis, and in any case this type ofpulmonary tuberculosis in adults is becoming rarer.Now let us approach the question from the opposite

angle. What percentage of known syphilitics developtuberculosis ? The high figures sometimes quoted fromBronfenbrenner (43 per cent.) and Tedeschi (70 per cent.)are open to various interpretations, and it is safer toquote more recent sets. Samson 9 examined a groupof I,300 prostitutes in Berlin and found amongst theWassermann-positive cases I2-5 per cent. of cases withactive tuberculosis, and amongst the Wassermann-nega-tive cases I0y7 per cent. Elliott 10 has put the figurelower, viz., 3 to 5 per cent. of tuberculous patientsamongst the syphilitics attending the University Hospital,Michigan.With regard to tuberculosis as a terminal complication

of late syphilitic and parasyphilitic diseases the evidenceis conflicting. Osler,11 on the authority of Stokes, statedthat tuberculosis was not uncommon in aneurysm, but inthe last 53 cases of thoracic aneurysm examined postmortem at Victoria Park, it has been found only twice, ararity comparable with the association of valvular diseaseand tuberculosis. Mott 12 stated that recent active

295

copyright. on F

ebruary 22, 2022 by guest. Protected by

http://sti.bmj.com

/B

r J Vener D

is: first published as 10.1136/sti.4.4.293 on 1 October 1928. D

ownloaded from

BRITISH JOURNAL OF VENEREAL DISEASES

tuberculosis was commonly found in autopsies on generalparalytics, especially females, but it appears to be fairlycommon also in dementia praecox and in acute maniacaland melancholic states, which are non-specific, whilst itis uncommon in tabes. It would be unjustifiable, there-fore, to attribute terminal tuberculosis to the syphilitictaint.

There is, however, one small but important group ofcases exhibiting the coincidence of syphilis and tuber-culosis in which the evidence of the effect of one diseaseupon the other seems indisputable, viz., the relativelyrare cases in which a patient with active phthisis acquiressyphilis. Here all writers seem to be agreed that thesecondary syphilis, with its accompanying systemicdisturbance, exercises a definitely bad effect on the activetuberculous lesion.

Records of post-morten examinations frequently donot refer to the association of these two diseases. Lands-berger 13*published a series of 5,332 autopsy records inwhich syphilis and florid tuberculosis were associated ino077 per cent. of cases and syphilis and healed tubercu-losis in og99 per cent.

Finally, there remains the question of inherited syphilisand tuberculosis. Here, on the one hand, we have thestatement of Caronia and Marinuccil4 that a series of70 children in Naples with bone and joint tubercu-losis revealed an incidence of 72-3 per cent. inheritedsyphilis, and another Italian record by de Angelis 15 tothe effect that tuberculin and Wassermann tests onchildren in an orphanage gave in one group a percentageof double infection as high as 48-5. Hutinel and Merk-len 16 also comment on the existence of inherited syphilisin tuberculous meningitis (ii out of 206). In thiscountry, Munro,16 in a careful piece of work, recordedii per cent. of inherited syphilis cases in tuberculouschildren in Fife and Kinross Sanatorium. These cases,all of whom had a positive Wassermann reaction, showedin all but one instance other stigmata of hereditarysyphilis. With regard to their tuberculous disease,either glandular or pulmonary lesions or both werepresent in all cases but one, which was a case of spinaldisease. Taking the incidence of hereditary syphilis ina general child population as being something underI0 per cent., Munro 17 considers from a study of his own

296

copyright. on F

ebruary 22, 2022 by guest. Protected by

http://sti.bmj.com

/B

r J Vener D

is: first published as 10.1136/sti.4.4.293 on 1 October 1928. D

ownloaded from

DIAGNOSIS OF TUBERCULOSIS

cases that syphilis as predisposing to a subsequent tuber-culous infection is not proven.What conclusions, therefore, can we draw from all

these records, many of them divergent ? Obviously,they must be very tentative.

(i) It is reasonable to suppose that the malnutritionso often accompanying inherited syphilis mayoffer a suitable soil for the tubercle bacillus, butthere are many difficulties in proving the point.

(2) Clinicians appear to be agreed that syphilisacquired during the active stage of pulmonarytuberculosis exercises a bad effect on the tubercu-losis.

(3) In the later stages of syphilis ulcerating surfacesmay, as pointed out by Sergent, offer a suitablepoint of entrance for the tubercle bacillus.Possibly an example of this effect may be seenin some cases of laryngeal tuberculosis.

(4) The diffuse so-called syphilitic fibrosis probablydoes not render a patient more liable to tuber-culosis. It may indeed protect him, since thetubercle bacillus does not readily attack fibroustissue.

(5) When large numbers of cases, including both latentand active, are analysed, the predisposing effectof the one disease upon the other is not wellmarked.

THE EXPERIMENTAL EVIDENCEThe spirochaete does not lend itself readily to labora-

tory experiments, especially in animals. It is strictlyanaerobic, whereas the tubercle bacillus is equallyobstinately aerobic.As regards symbiosis little is known. A few cases have

been recorded in which both parasites have been foundtogether in the lymph nodes of man. W. and R. Spitzer 18examined three such cases, two of fistulous cervicaladenitis, the third of caseous epitrochlear nodes with ageneralised papular syphilide. Spirochietes were foundin all three cases, and all three glands produced genera-lised tuberculosis when inoculated into guinea-pigs.The same workers have attempted to graft syphilitic

infection on to rabbits previously injected with bovine297

copyright. on F

ebruary 22, 2022 by guest. Protected by

http://sti.bmj.com

/B

r J Vener D

is: first published as 10.1136/sti.4.4.293 on 1 October 1928. D

ownloaded from

BRITISH JOURNAL OF VENEREAL DISEASES

tubercle bacilli, and vice versa, but without any apparentinfluence upon the course of either disease. Unfortu-nately, the strain of bovine tubercle bacillus used wasfound to be of low virulence.Numerous allergy experiments have also been made,

but they are not very convincing, and the attempt toseparate tuberculosis into three stages, primary, secon-dary and tertiary, like syphilis, is rather arbitrary.Most of the work (Dujardin and Duprez 19) on allergyhas been done with skin tests, using leutin and tuberculinas the two antigens, but this method is by no means freefrom error. Perhaps the most interesting result (Lelongand Rivalier 20) is that of a positive tuberculin reactiondiminishing in intensity with the onset of secondarysyphilis.

Further animal experiments are needed, but they areobviously difficult to make, owing to the refractorycharacter of the spirochaete.

DIAGNOSIS

First, as to serological methods. The Wassermann testis now so well established that a discussion on it isunnecessary here. The chief difficulty is the point raisedby some workers as to the possibility of non-specific crossfixation. This question has been investigated by severalworkers. Dulaney 21 made a careful test of 6oo sera,ioo being sera of tuberculous patients and 500 of routineblood samples sent for Wassermann tests. Eight of the I00sera from tuberculous patients gave a positive Wassermannwith cholesterolised antigen, but this proportion is notnotably higher than results obtained by other workers withnon-cholesterolised antigen. In the group of 500 routineWassermann sera, I5 from patients who had no signsof tuberculosis gave positive fixations with both tubercu-losis and Wassermann antigens, i.e., 3 per cent. Theseresults appear to me to be more striking than those in thefirst group, but it must be remembered that activetuberculosis is sometimes difficult to diagnose clinically,and that a positive complement-fixation test M& ithoutclinical signs is not necessarily incorrect. The conclusionappears to be that (i) positive Wassermann tests occur-ring in known tuberculous patients should be repeatedwith both cholesterolised and non-cholesterolised anti-

298

copyright. on F

ebruary 22, 2022 by guest. Protected by

http://sti.bmj.com

/B

r J Vener D

is: first published as 10.1136/sti.4.4.293 on 1 October 1928. D

ownloaded from

DIAGNOSIS OF TUBERCULOSIS

gens; (2) positive complement-fixation tests in syphiliticsshould l)e accepted with reserve. If I may recordwith own experience, I had encountered no difficultvmy cholesterolised antigen until we came to investigate,at Nictoria Park, a special group of mal-nourishedso-called" pre-tuberculous" children. Here the questionof inherited syphilis arises in a particularly diffi-cult form, and the cases are being tested with bothcholesterolised and non -cholesterolised antigens. Thisinvestigation is still going on. I have no conclusions tooffer as yet, but I am beginning to doubt if the presenceof cholesterol is a serious drawback.

The differential diagnosis in tissues may offer consider-able difficulties. A few cases have been recorded inwhich both Tr. pallidum and B. tuberculosis have beenfound in the same lymph gland. Similarly, Sergentstates that B. tuberculosis has been found on primarysores, but if so, such a happening must be extremely rare,and one would always be afraid that confusion with thesmegma bacillus had occurred unless an animal inocula-tion test had been made.

In cases in which it is not possible to make a Wasser-mann test reliance must be placed on other diagnosticdistinctions. It is noteworthy that syphilis tends toaffect the cardio-vascular system whilst tuberculosis does1not. Syphilitic lesions of the liver are common; in tuber-culosis lesions in the liver are rare. On the other hand, tnetubercle bacillus readily finds its nidus in the lung, wherethe spirochaete is only very rarely found. Enlargementof lymph glands, with frequent caseation, fills the picturemuch more in tuberculosis than in syphilis, where breakingdown is rare. These differences could be multiplied to aconsiderable extent. It would appear, indeed, that thereis a deep underlying difference between the cellularresponses of the tissues to these two parasites, the one abacterium, the other a protozoon, and we see this differencestill further stressed in the effect of antiseptics andchemotherapeutic remedies on the two parasites. Inmorbid histology the real difficulty is likely to arise in thecase of small gummata. The prominence of new-formedvessels with scanty formation of epithelioid and giant cellsin syphilis, and the presence of avascular caseation withgiant cells, many epithelioid cells and even tuberclebacilli in tuberculosis are the chief points of distinction.

299

copyright. on F

ebruary 22, 2022 by guest. Protected by

http://sti.bmj.com

/B

r J Vener D

is: first published as 10.1136/sti.4.4.293 on 1 October 1928. D

ownloaded from

BRITISH JOURNAL OF VENEREAL DISEASES

Too much stress cannot be laid on the finding of thetubercle bacillus. It is all-important. Ziehl Neelsenstained sections are a necessity. In other words, histo-logy is of vital importance in tuberculosis, but of secondaryimportance in syphilis.

Syphilis of the lung offers special difficulties, Threeforms have been described: (i) bronchitis during secon-dary syphilis; (2) gummata, and (3) diffuse pulmonaryfibrosis in the tertiary. The first mentioned offers littledifficulty. Gummata of the lung I have not seen inseventeen years' experience of post-mortem work atVictoria Park. Pulmonary fibrosis with a positiveWassermann is not very uncommon, but, in the absenceof other signs of syphilis, it is by no means certain that thefibrosis can be regarded as of specific origin, since fibrosis.of lung without syphilis is a fairly common disease. Indoubtful cases the sputum should be examined for tuberclebacilli daily for at least a week before being passed asnegative, and in children the faeces should also beexamined for tubercle bacilli. In passing it may be notedthat in a recent series of cases I have found mouthspirochoetes in IO per cent. of sputa sent for routineexamination for tubercle bacilli. Broncho-pulmonaryspirochoetosis, a disease in which haemoptysis is a promi-nent symptom, seems to be a separate entity and hasprobably no relation either to syphilis or tuberculosis.The so-called lymphosarcomata appear occasionally tohave been regarded as gummata. It must always beborne in mind that a positive Wassermann reaction mayoccur with either malignant disease or tuberculosis of thelung. Gummata of the sternum may complicate thediagnosis occasionally.The examination of the cerebrospinal fluid for the

purpose of differentiating between tuberculous meningitisand syphilis of the central nervous system can only bementioned briefly in passing. As a general rule, thetubercle bacillus can be found in the fine skein of clotwhich forms in the fluid soon after withdrawal, whilst theWassermann reaction is positive in the syphilitic caseseither in the blood or cerebrospinal fluid. The cellcount, globulin and protein estimations, and the colloidalgold tests are subsidiary tests which need not be dealtwith in detail. They are chiefly important in support-ing the Wassermann test.

300

copyright. on F

ebruary 22, 2022 by guest. Protected by

http://sti.bmj.com

/B

r J Vener D

is: first published as 10.1136/sti.4.4.293 on 1 October 1928. D

ownloaded from

DIAGNOSIS OF TUBERCULOSIS

Perhaps one of the most difficult groups of cases inwhich to distinguish syphilis from tuberculosis is thesurgical tuberculosis group of bone and joint diseases,especially when there is no discharging sinus fromwhich to obtain pus for examination. A positive Was-sermann does not necessarily exclude tuberculosis,and when possible, any tissue available should be sub-mitted to histological examination. The same remarkapplies to cutaneous tuberculosis, especially in Bazin'sdisease.

Finally, there are the questions of the interpretationof the complement-fixatfion test in tuberculosis and of thevarious flocculation and precipitation tests in bothdiseases. The value of the complement-fixation testdepends largely on the choice of antigen. For my ownpart, I much prefer the use of antigens consisting oftubercle bacilli suspended in saline rather than thosemade from liquid cultures containing egg, etc. Diffi-culties of cross-fixation arise here, as in the case of theWassermann test, since tubercle bacilli possess a con-siderable amount of lipoid substances. On the whole itseems safe to regard a definitely positive reaction asindicating active tuberculosis, but it is by no means clearhow a weak positive result should be interpreted, especiallyif the patient should happen to give a positiveWassermann.In all cases a graduated series of dilutions of serum shouldbe used, as in making Wassermann tests on cerebrospinalfluid.A large number of precipitation and flocculation tests

have been devised both for tuberculosis and for syphilis.I have no experience of them in syphilis, but in tuberclethey are one and all unreliable.We are, indeed, greatly in need of a reliable serological

test for tuberculosis. Nowadays new serum tests for thisdisease appear with conspicuous regularity only to befound wanting and relegated to the literature of the past.The sedimentation test appears to be the best of the non-specific ones, but I do not know how far it could be usedas a means of differential diagnosis between syphilis andtuberculosis. I have refrained from discussing thetuberculin tests because I do not regard them as goodtests of the activity of tuberculous disease; moreover,tuberculin at present is difficult to standardise withcertainty.

30I

copyright. on F

ebruary 22, 2022 by guest. Protected by

http://sti.bmj.com

/B

r J Vener D

is: first published as 10.1136/sti.4.4.293 on 1 October 1928. D

ownloaded from

BRITISH JOURNAL OF VENEREAL DISEASES

REFERENCES(I) L. HOLLANDER and F. C. NARR. Archives of Dermat. and

Sykphilol., I92I, IV., I52.(2) E. V. ADELUNG. Journ. Amer. Med. Assoc., I9I8, LXX., I2II,(3) A. K. BOWMAN. Lancet, I923, II., I288.(4) R. A. KILDUFFE. Archives ofDermat. and Syphilol., I922, V., 207.(5) J. RITTER. Die Therapie de Gegenwart, I927, LXVIII., 247

and 300.(6) D. POWELL and H. S. HARTLEY. Diseases of the Lungs and

Pleurae." I92I.(7i E. SERGENT, quoted by A. MIRANDE. "Tuberc1e," I9I9,I., 1o0.(8) E.MORTON. "Tubercle,"I920,II.,I50.(9) J. W. SAMSON. Zeitschr. f. Tub. 1923, XXXIX., i6i, Quoted

in Amer. Rev. Tub., I925, XI., 57.(io) ELLIOTT.(II) W. OSLER. ALLBUTT and ROLLESTON'S System of Medicine,"

Vol. VI., I909.(I2) F. W. MoTT. Archives of Neurology, Vol. IV.(I3) LANDSBERGER. Virchows Arch. f. Pathol. Aust. u. Physiol.,

24I, 392, quoted by H. LOSSEN, Beitrage fur Klinik der Tuberculose,1927, LXVI., 76I.

(14) G. CARONIA and R. K. MARINUCCI. Pediatria, I9I7, XXV.,7I3; abstracted Amer. Rev. Tub., I9I8, II., I83.

(I5) F. DE ANGELIS. Pediatria, I923, XXXI., 357; abstractedAmer. Rev. Tub., I923, VIII., IIO.

(I6) F. HUTINEL and P. MERKLEN. Arch. de Med. d'Enfants, I92I,'XXIV., 52I; abstracted Amer. Rev. Tub., I92I, V., 227.

(I7) W. T. MUNRO. Lancet, I924, II., II73.(i8) W. and R. SPITZER. Klin. Wochenschr., 1922, I., I5; ab-

stracted in Amer. Rev. Tub., I922, VI., 208.(I9) B. DUJARDIN and C. DUPREZ. Rev. Belg. d. 1. Tuberc., I924,

No. 3, 87; abstracted Amer. Rev. Tub., I924, X., 58.(20) M. LELONG and E. RIVALIER. Compt. Rend. Soc. de Biol.,

1923, LXXXVIII., 327; abstracted in Amer. Rev. Tub., I923,VIII., 26.

(2I) A. D. DULANEY. Amzer. Rev. Tub., 1922, VI., I92.

302

copyright. on F

ebruary 22, 2022 by guest. Protected by

http://sti.bmj.com

/B

r J Vener D

is: first published as 10.1136/sti.4.4.293 on 1 October 1928. D

ownloaded from