THE BACTERIOLOGY OF EPIDEMIC INFLUENZA WITH A DISCUSSION OF B. INFLUENZÆ AS THE CAUSE OF THIS AND OTHER INFECTIVE PROCESSES

By W. L. Holman, B. A., M. D.

Introduction

In a study of the bacteriology of a respiratory disease such as influenza, the technical difficulties encountered are very great and must be overcome before we can draw useful conclusions from the results obtained or attempt to determine the etiological factors. The important methods of attacking such a problem include: (1) the study of stained smears and cultures from the various available materials, along with the demonstration of the bacteria in the lesions found in the disease by a study of sections; (2) tests with the various materials to determine the presence of the causative agent, which includes experiments on man and animals and is more inclusive than the mere study of the bacteria isolated; (3) immunological studies of man suffering from the disease, or of man and animals treated with the materials from the disease; (4) pathological, clinical and epidemiological studies linked with the above.

Many of the difficulties and sources of error in these methods are manifest to all, but certain points may be indicated as more important in the phases of the work on which I am to report.

General Methods of Investigation

Stained smears from the material available. The choice of the material is of first importance. Sputum to be of any real value must be obtained from the deeper portions of the respiratory tract, should be as free as possible from the secretions of the buccal cavity, and should be washed in saline before it is used. These are considered among the first requirements in the study of lung infections by the pneumococci and are equally important in influenza. Swabs from the nasopharynx should be obtained with the same precautions as are demanded in meningococcal work. The other available material—such as blood, lung puncture fluid, pleural fluid and spinal fluid—must be collected with the greatest care.

The staining methods should, naturally, include those which will bring out the various types of bacteria, and must include the Gram method, using dilute alcoholic fuchsin (1-20) as the counterstain. The varying morphology of the B. influenzæ and its frequent minute size make it difficult to detect. It is not the only Gram negative small bacillus seen in smears from the throat, but when it occurs in the typical schools, or where there are numerous bacilli to be seen, its characteristics are quite definite. I have recently isolated an anærobic Gram negative bacillus from a series of swabs from the buccal cavity which suggests in many ways the morphology of the B. influenzæ, which will indicate one of the many difficulties to be met with in the study of stained smears. They are, nevertheless, of great use as a control on cultures, and most helpful in the study of the material from sources other than the respiratory tract.

Cultures of the bacteria from the various materials. Here we have the greatest difficulty of all. The medium chosen determines the bacteria which will appear to predominate, and there is no single medium that will answer all purposes. Streptococci will appear to be in excess when serum broth is used, as I have previously shown; pneumococci with Avery’s pneumococcus medium; and staphylococci, the Gram negative cocci, and the diphtheria group with Loeffler’s serum. Ordinary blood agar is perhaps the best general medium for direct and secondary plating. There have been many special media devised for growing the B. influenzæ, but the one I have used most and found particularly helpful is heated blood agar made after the general method of Voges.

The extremely tiny colony of B. influenzæ on ordinary blood agar makes it particularly difficult to detect, and one is apt to get the wrong impression of its numbers from the macroscopic appearance of the plate. In attempts at isolation there must be a liberal use of media in picking colonies, as many suspicious ones will turn out to be immature growths of B. xerosis, M. pharyngis (or M. catarrhalis), streptococci, or more rarely pneumococci and other organisms. Replating from such picks is frequently necessary, and further plates, from the original culture on heated blood agar, must often be made before the B. influenzæ can be isolated. The care required in all stages of the isolation of this organism, the unstinted use of media for plating and for picks, the number of stained smears to be studied, and the further transfers necessary to verify results, all these limit the amount of material which can be studied with any degree of accuracy. If further the streptococci, the pneumococci, the Gram negative cocci, the capsulated Gram negative bacilli and many others are to receive any attention, it can readily be appreciated that a few cases carefully studied are of far more value than a large number hurriedly examined in an uncertain routine.

The pathological study of the same cases on which I have done the bacteriology will be found in Dr. Klotz’s paper in these communications, and I will merely refer to some of the bacterial findings in the sections of the lungs and bronchi. The more inclusive methods which have been used in attempts to determine the etiological factor in influenza we have been unable to attempt, but I will refer later in this paper to the findings of the investigations of others. Immunological studies have been limited to a few investigations on the presence of agglutinins, complement binding substance, skin reactions and the amount of complement present in the sera of certain patients. The epidemiological and clinical studies are reported by Drs. Johnston and Lichty in this series of reports.

Material Studied

The material used in the study I am reporting included swabs from the large bronchi and fluid from the lungs and pleural cavities of 32 autopsies, as well as blood cultures from 22 patients and swabs from the nasopharynx of 31 individuals. Fifteen sera were tested for fixation of complement with an antigen made from several strains of B. influenzæ. Fourteen other sera were tested for agglutinins. Complement content was determined in the sera of 25 patients. Skin tests after the Von Pirquet method were done on 14 convalescents, and carefully stained nasopharyngeal smears without cultures were studied from 48 patients.

The chief attention was given to the study of the autopsy material and we concentrated on the isolation of B. influenzæ. At the same time we did not neglect the other bacteria making up the flora of the bronchi, lungs and pleural cavity in these cases. The various types were isolated and most of them fully identified.

Technique

Direct smears were made on sterile slides of all material studied and stained by Gram’s method. The counterstain was always alcoholic fuchsin diluted 1-20 in distilled water. Direct cultures were made on a human blood agar plate containing 5 per cent. blood, which was further smeared just before use with defibrinated blood. This latter procedure was later discarded, as it did not appear to assist to any marked extent the growth of B. influenzæ. Blood broth containing a few drops of defibrinated blood and blood agar slants smeared with blood were also used. Heated blood agar (2-3 c.cm. of defibrinated human blood added to 100 c.cm. of ordinary agar at a temperature of from 90 to 100° C., or as the agar comes from the sterilizer) was used in the last nine cases to replace the blood agar slant in the direct cultures and as the medium of choice for transfers of the B. influenzæ.

I prefer the ordinary blood agar plate to the heated blood plate because the former gives readings which are very helpful in distinguishing colonies of various types. B. influenzæ appears as clear, tiny, pinpoint, inert colonies. B. xerosis or the pseudodiphtheria group gives more opaque but often rather similar colonies. Gram negative cocci as M. pharyngis siccus have dry, raised, soon becoming wrinkled, inert colonies, varying greatly in size; M. catarrhalis, more moist, inert colonies. The cocci of the streptococcus viridans group appear as very small colonies with greening, or are not infrequently inert, while thin, flattened colonies with central thickening may sometimes be noted. Those of the streptococcus hemolyticus group occur as small, frequently nipple-like colonies with clear, wide zones of hemolysis; pneumococci as moderately small, moist, dewdrop-like colonies with center collapsing early and with greening; streptococcus or pneumococcus mucosus as larger, watery, sticky colonies with greening and frequently an early clearing near the colonies.

TABLE I.

BACTERIOLOGY OF THIRTY-TWO AUTOPSIES FROM INFLUENZA CASES.

Autopsy Number.Date.Day of Disease.Hours P.M.DIRECT SMEAR—GRAM’S METHOD.B. InfluenzæPneumcocci.Strept. Mococci.Hemolytic Strept.S.P.A.Other Cocci.Other G—B.Other Bacteria.NOTES.
Bronch.Lung.Pleural Fluid.
7411918 Oct. 9316G +staph. Few pneumo-like. Few chains of elong. cocci.0+0 +G+ diploc. Nine plates used to isolate B.I. Sp.a. overgrew all cultures. B.I. seen in blood smear agar in 24 hours.
7431158Br. G—bac. from coccoid to short threads. Mostly scattered. Some phagocyted. Fewer G +cooci in short chains.+0– Pleural fluid and seen as diplos in direct smear. Br. G + diploc. not like pneumo.Br. lux. white almost coccoid. Pericard, fluid and liver juice, no growth.
74411711Br. G—bac. moderately stout about in small groups and scattered. G+diploc (pneumo) also G— threads. Phago. of both in a few cells.+–0Lung + M. tetrag. in Br. M. pharyng. in Br. Pneumococcus from lung. No attempt after first plate to isolate B.I.
74512106Br. G +–large bac., strept. short, G—B, few, very short, no threads.+0– Pleural fluid, also seen in smear.Pl. fluid, also seen in smears. Swab from ruptured rectus. Sterile. No material from lung.
746125½Br. G—B very short, no threads. Irregularly scattered. More seen in left bronchus. A few cells phagocyted.+–0 Strept. viridans from bronchus.B. coli from bronchi and lung.B. xerosis from bronchus.The overgrowth of B. coli in lung material prevented further attempts to isolate B.I.
7471363Br. G+diploc, fairly numerous. G—B tiny, as diplos and in long threads scattered or in small groups. Pleural fluid and lung no bacteria seen.+–– Bronchus and pleural fluid.Strept. viridans from bronchi and lung. B. subtilis group from pleural fluid.Five picks from blood agar plate failed to recover B.I. from lung.
7481344Br. nothing like B. I. seen. G+ small elong. diplo. Numerous G + diploc. in lung. Comparatively few Q-B, very short.–+0Lung+, not isolated from bronchus. Strept. viridans from bronchus. B.I. not seen nor isolated from the bronchi.
74914415Br. G+large pneumo like, many G+large bacilli, single and in pairs. Few G—B very tiny and widely scattered; lung, heavy mixture as in bronchi.––0Bronchus Lung? Bronchus and lung. B. coli from bronchus and lung. The overgrowth of B. coli prevented any further attempts to isolate B.I.
7501496Br. G+B large like B. welchii, G—rather stout coccoid forms, G+C in pairs and short chains. Tiny G—coccoid forms like B. I. Lung G+ pneumo-like and caps, chains; no B. I.––0Bronchus? Lung? B. coli from bronchi and lungs. B. coli again present as in No. 749. Direct smear suggests heavy contamination.
7511476Br. G +cocci large elong.? caps, also G +C in flat pairs. G—coccoid forms. Lung, numerous bacteria. G+strept. with flattened cocci. Some G-short forms?––+ Pleura. Lung. Bronchus. M. tetragenous from bronchus. Spore-bearer with tiny cols, pleur. B. xerosis from bron.
752151315Br. G+pneumo-like. G+B smaller than B. welchii, occasionally tiny G -diplobacillus. Lung, G+chains of cocci Gram weak. Few G—tiny bacilli scattered or in groups.–+0Bronchus and lung. Strep. viridans from bronchus and lungs. B.I. like seen in original culture on blood agar but not isolated from bronchus.
75616818Br. numerous G+B. B welchii like. G—B large and few tiny. G+round diploc. Pl. fluid almost pure pneumo-like, few G-forms probably the same.–0– Pleural fluid.Strep. viridans from bronchus.B. coli from bronchus and pleural fluid.B. xerosis from bronchus.Compare No. 749 and 750. Fluid from lung not obtained for culture.
75716614Br. G—B tiny, to medium. G—like M. catarrhalis. G+cocci, pairs and chains. Few B. W. like. Lung, many G—B like B. I. Some cells filled, also G—cocci. M. catarrhalis like and rare B. welchii like.++0Bronchus and lung. Bronchus and lung.M. tetragenous? from lung. B. xerosis from bronchus.This case 14 hours P. M. gave B.I. from all the material.
758161416Br. pneumo-like in excess. G—B from tiny to forms stouter than B. I. Few strept. rare M. catarrhalis. Lung, pneumo-like. Phago.+?0Bronchus and lung. M. catarrhalia-like from lung. Strep. viridans from lung and bronchus.B. coli from bronchus. The B. coli did not prevent the isolation of B.I. like seen in original blood agar cultures of lung.
76117719Br. pneumo-like. B. I. like common, M. catarrhalis like. Both B.I. and M. catarrhalis phagocyted. B.I. single or in threads. Some typical groups. Lung, pneumo, caps, rare, M. catarrhalis like.+–0Bronchus and lung. Bronchus and lung. B. coli from bronchus. Even after 19 hours P. M. the B.I. was isolated.
762171012Br. numerous B.l. like typical, also many pneumo. and M. catarrh. Lung same. M. catarrh. phagocyted. B.I. smear, many phagocyted, many pneumo.+++Pleural fluid and bronchus. M. catarrh. like from lung and bronchus. B. xerosis from lung. B. subtilis from bronchus.
763171113Lung, pneumo-like, slight phagocytosis. Pl. fl., pneumo and few strept., slight phagocytosis.0––Pleural fluid. No growth from lung on plate. B.I. like seen in original culture from pleural fluid. No material from bronchus.
7641796Br. B.I. smear. Cells crowded. Pneumo-like fewer, occasional G—stouter thread.+00Bronchus. Staph, albus from bronchus. Material only from bronchi.
76517916Br. pneumo. B.I. few scattered. G+flattened diploc. Phago. of B.I. and pneumo. Lung, pneumo-like, rare strept. very questionable G—B free and in cells.++0 Lung. M. catarrh. from bronchus and lung.B. coli from bronchus and lung.
767181014Br. rather round pneumo-like with caps. B.I. few. Scattered, also in cells. Lung, few bacteria. G+strep. often phagocyted.––0Bronchus.Lung. Bronchus. Blood culture 15
10 gave pure growth of pneumo. mucosus.
77019119Br. crowded with B.I. like. Few G+cocci and fewer M. catarrh. like. Pl. fluid G+flattened pairs, pus cells, phagocyted.++–Bronchus. Bronchus, lung, pleural fluid.
7732120 Recurrence.3Br. few bacteria G+and G—pneumo-like. Rare G+–thread. Lung, pneumo and rare strept. Pl. fluid, pneumo-oat shapes, etc.––– Bronchus.Strept. viridans bronchus. Sarcina albus lung. B. xerosis from bronchus. G + B lux. white pleura. fluid.No growth from lung except sarcina. Only 2 colonies from pleural fluid on blood agar plates.
778242317Br. B.I. smear. Fewer large pneumo. Lung, G + small diploc. Few B.I. like. Pl. fluid, few cells, no bacteria.++–Bronchus and lung. Lung. Non-motile, non-fermenting, lux, white from bron. Ten plates and 30 picks were done for the isolation of B.I.
7812654Br. crowded with staph. like. Fewer G—B, larger than B.I., few M. catarrhalis like. Lung G+ small staph. like, caps, cocci in pairs and chains. Few tiny G—B. Pl. fluid pneumo-like and elong. cocci in chains capsulated.–+– Lung and pleural fluid. Bronchus and lung abscess.Staph. albus and sarcina from pleural fluid. B.I. like seen from 24 hour Ht. blood agar from bronchi and lung but only isolated from lung on replating. Bl. culture 25
10 sterile.
7822683Br. numerous B.I. like scattered, some phagocyted. Fewer G+ flat pairs with capsule.+–0 Bronchus and lung. No B.I. like on 24-hour Ht. blood agar from lung.
7832681Br. G+small caps, pneumo-like. Lung poor smear, occasional pneumo-like.+–0Lung. Bronchus.M. catarrh. like bronchus. No B.I. like on 24-hour Ht. blood agar from lung.
7842886Br. capsulated pneumo-like, few strep. Lung, chiefly pneumo-like. few G—B like B.I., also G—pneumo-like.++0Bronchus and lung.Bronchus and lung? M. catarrh. like bronchus. Numerous B.I. like on 24-hour Ht. blood agar of bronchi and fewer from lung. Isolated by replating.
7862942Br. G+cocci in round pairs and rather flat chains, suggested caps. Tiny G—B very rare. Lung streptococci flattened, often phagocyted.+–0Bronchus. Bronchus and lung. Staph, albus from bronchus. Pleural fluid not collected sterilly, Haemol. strept. isolated.
7872982Br. numerous pneumo-like, bacillary forms. A rare suspicious B.I. like, some of these in cells. Lung, caps, elongated diplos, and chains of elong. cocci.++0 Bronchus and lung. B.M.C. from bronchi. All the bacteria isolated were seen in 24-hour Ht. blood agar cultures from bronchi and lung.
791Nov. 166Br. few bacteria. G+pneumo-like round, G—B and threads, size varies, like B.I. Lung, G + caps, pneumo. G+Large B. few suspicious G—coccoid forms. Pl. fl. caps, pneumo and caps, elong. chains.++–Bronchus and pleural cavity. Bronchus and lung. Replated from Ht. blood agar to isolate B.I. from lung.
792263Br. caps, pneumo-like bac. forms and chains. G-caps, pneumo-like. Few G—B. questionable. Lung. caps, pairs and chains of elong. cocci, in cells. Pl. fluid, numerous caps, chains of diploc.++–Bronchus, lung and pleural fluid. Staph. albus, lung, strept. viridans lung, M. catarrh. like lung and bronchi. B.I. like seen on 24-hour Ht. blood agar from bronchi and lung but not pleural fluid.
7934103
2		Br. M. catarrh. and G+cocci, few bacteria, few G—B. Ear, G+cocci.–––?Throat. Bronchus, lung, arm vein, spleen ear.Throat, ear and bronchus.Strept. viridans from throat.B. coli from throat. B.I. like never seen except from throat which may have been B. coli.
Total20132206416
B. influenzæ found—Percentage66½4614
Total percentage for B. influenzæ 78%

EXPLANATORY NOTE. B.I.—B. influenzæ. S.P.A.—Staphylococcus pyogenes aureus. M. pharyog—Micrococcus pharyngis siccus. Br.—Bronchus. Phago.—phagocytosis. Ht.-Heated blood agar. B. W.—B. welchii.

Staphylococci develop opaque, paint-like colonies of varying size, with or without hemolysis, and so do other less frequently found bacteria give more or less distinctive colonies. The heated blood agar does not show these differences.

The colonies most liable to be confused with those of B. influenzæ are, therefore, those of B. xerosis, immature colonies of the Gram negative cocci and certain colonies of the streptococcus viridans group. Transfers should always be made to heated blood agar of all colonies suggestive of B. influenzæ, or when the growth of the B. influenzæ has only occurred in the more crowded portions of the plate, and it is difficult to pick pure cultures, attempted pickings should be made to this medium for further platings. It is frequently necessary to make further blood agar plates from the original blood agar, blood broth or heated blood agar cultures after longer incubation periods, depending on the findings in smears from these media. The heated blood agar is the best of these to encourage the growth of B. influenzæ. It must, however, be used at once, or within a very few days of its preparation, and cannot be kept on hand as a stock medium. I have not found it as useful for plating because of the difficulty of differentiating colonies. The phenomenon of the star-like and more luxuriant growth of the colonies of B. influenzæ about colonies of other bacteria has often been noted, and will be referred to in a later portion of this report. Here it may be said that this is at times a marked feature of certain mixtures and must be recognized in studying the plates. The finding of B. influenzæ in picks from apparently isolated colonies of other forms is not uncommon, and is the same type of difficulty which I have discussed in papers on streptococci. It is important to recall, in connection with cultures taken from the lungs at autopsy, the experimental work of Norris and Pappenheimer, who showed that B. prodigiosus put in the mouth immediately after death could be recovered from the lungs in over 50 per cent. of the cases studied.

Results of the Author

In Table I are shown my results from the 32 cases which came to autopsy. The B. influenzæ was isolated from one or more sources in 25, making a total of 78 per cent. Most of the negative cases probably also had this organism, but I did not grow it from the material which I used for culturing. The work of others would indicate that it may have been present in other regions, such as the sinuses of the head or other portions of the lung and respiratory tract. The positive results show B. influenzæ present in 20 out of 30 cases from the bronchi; in 13 of 28 from the lungs; in 2 of 14 from the pleural cavity; in 9 of 26 from both bronchi and lung where both were cultured; in 8 of 26 from the bronchi with the lung negative; in 3 of 26 from the lung with the bronchi negative; once of 10 from the pleural cavity with both the bronchi and the lung negative, and once from all three sources.

The negative results occurred in seven cases. In three of these (749, 750, 756) B. coli overgrew the cultures from the bronchus, in two also from the lung, and in one, without lung culture, from bronchus and pleural cavity. The mere presence of B. coli, however, did not preclude the isolation of B. influenzæ, as is seen in cases 746, 758, 761 and 765. The finding of B. coli would suggest a post-mortem invasion. The hours after death before the autopsy was done were in these seven cases, ½, 15, 6, 18, 16, 19, 16, respectively. That delay in performing the autopsy, as emphasized by Spooner, Scott and Heath, adds to the difficulty is self-evident, but successful isolations of B. influenzæ have been obtained after even longer periods than in the negative cases (761). In the fourth negative case (763) the bronchus was not cultured. A pneumococcus was grown from the pleural cavity and no growth was obtained from the lung. In the original culture from the pleural cavity influenza-like forms were seen but could not be isolated. In the fifth case (767) a blood culture three days before death gave a growth of pneumococcus mucosus which was also grown from the lung at autopsy. Direct smear from the bronchus showed very few influenza-like forms. Our sixth negative finding was in a case of 20 days’ illness, the patient having had a recurrence (773). Staphylococcus pyogenes aureus, streptococcus viridans and B. xerosis were grown from the bronchus. Only a sarcina form grew from the lung, and a further probable air contamination occurred on the media from the cultures of the pleural cavity. The B. xerosis colonies were confusing, picked as possible B. influenzæ, and, before this was discovered, the overgrowth prevented further attempts to isolate the influenza bacilli. The last unsuccessful case was one with a general infection of a hemolytic streptococcus from an acute otitis media. The streptococcus was isolated from the bronchus, lung, spleen, arm vein and the middle ear at autopsy.

It will be seen that in these seven negative cases technical difficulties prevented the isolation of the B. influenzæ, even if it had been present. I would not, therefore, conclude that the organisms were necessarily absent, but rather that we have failed either to secure material from the focus of infection or on account of the other reasons mentioned.

It is very evident that a variety of secondary organisms very frequently overgrow the field and become numerically predominant. In our first case staphylococcus pyogenes aureus overgrew all the other organisms present in cultures from the lung material. B. influenzæ was, however, seen in the original 24-hour blood agar culture. It required 9 blood agar plates before the organism could be isolated. In another case 10 plates were used for the isolation.

The findings of the bacteria in the lung sections are particularly interesting and instructive. The entire series of cases have not been completely studied, so I am unable to tabulate the findings. In cases 761 and 762 sections of the lung showed influenza-like bacilli to be almost pure in the earlier stages of the process, while in areas with purulent foci pneumococcus-like and other Gram positive cocci were also numerous. In some cases B. influenzæ-like organisms were to be seen in overwhelming numbers. In others they were scarce, while in some nothing resembling B. influenzæ could be found in the sections. Positive cultures were often independent of whether the influenza-like forms were to be seen in smears or sections or not, although they were found in the great majority of the cases. The findings in the direct smears and the bacteriological results make useful material for comparison.

Swabs from the nasopharynx were cultured from 31 individuals; nearly all of these were cases suspected of diphtheria or as carrying the diphtheria bacillus, and no particular effort was made to isolate the B. influenzæ. They were seen in the mixed culture occasionally. In the last eight cases the heated blood agar, ordinary blood agar and Loeffler’s serum were seeded from the throat swabs. B. influenzæ practically overgrew all the other bacteria from seven of these cases on the heated blood agar medium and was isolated without difficulty; all eight showed M. catarrhalis. The two other media gave little or no evidence of the presence of B. influenzæ. As I have said above, our attention was concentrated on the autopsy material. These cultures from the throat were simply made to demonstrate the usefulness of the heated blood agar.

TABLE II

BACTERIA SEEN IN DIRECT SMEARS FROM NASOPHARYNX

Type of Disease.Number of Patients.B. Influenzæ-Like.Pneumococcus-Like.M. Catarrhalis-Like.
Early2414176
Serious1313139
Convalescent118116
Total48354121
Percentage of positives 738643

Direct Smears from Nasopharyngeal Swabs

It is recognized by most of the modern investigators that little reliance can be put on the finding of B. influenzæ-like bacilli in direct smears. The organism is markedly pleomorphic, occurring as extremely small coccoid forms up to threads of various lengths. Notwithstanding these morphological variations the organisms are usually seen as tiny bacilli, and these are considered as the typical form. We carried out a series of microscopical examinations of carefully made smears from the throats of patients with influenza. Particular attention was given to the occurrence of organisms resembling in morphology and staining B. influenzæ, pneumococci and M. catarrhalis. We have divided the cases roughly into three types—early, serious, and convalescent. Table II shows our results. The term B. influenzæ-like was used for the typical morphological picture so often described. Dr. Frost and Mr. Scott carried out this portion of our work and their results are interesting.

Blood cultures were done on 22 cases. Pneumococcus mucosus was grown from one patient who three days later came to autopsy (Case 767). In another case pneumococcus-like organisms were seen in smears from the dextrose broth flask after 24 hours’ incubation. These, for some unknown reason, did not grow on blood agar plates. After 48 hours smears made on blood agar from the original flask gave a growth of B. influenzæ and a M. catarrhalis-like organism. I consider this result a very unsatisfactory one, being quite unable to explain the failure to grow the pneumococci-like forms on transfer. Possibly the acidity developed might account for it.

TABLE III

AGGLUTINATION TESTS WITH SERA OF CONVALESCENT INFLUENZA PATIENTS

Dilution of Serum++––
Convalescents1-1302
1-10527
1-40239
1-800113
1-1600014
Normal Controls1-10120
1-40012

The complete agglutination as would be indicated by +++ or ++ was not seen.

Agglutination tests were carried out with the sera of 14 convalescents and 3 normal individuals. A polyvalent emulsion of strains of the influenza bacillus isolated from our cases was used. The results are shown in Table III. Tubes were incubated at 37.5° C. The results did not indicate anything in the nature of a specific reaction. Dr. Frost carried out this work during the height of the epidemic, but we were unable to continue it further. A short review of the work of others will be found near the end of this paper. Miss Thompson and Mr. Mock studied complement fixation, using the sera of 15 convalescents against an antigen of B. influenzæ. Their results were negative. The antigen appeared to be slightly more anti-complementary than were emulsions of staphylococcus or B. coli. Huntoon also noted this anti-complementary character of emulsions of B. influenzæ.

Attempts were made to estimate the amount of complement present in the fresh blood serum of influenza patients. The technique was to use a 1-4 dilution of the patient’s serum, adding measured amounts of this to a 1 per cent. blood emulsion, with 1 unit of amboceptor and determine the smallest amount necessary to bring about complete hemolysis. This test was carried out on eight patients ill for only a few days. The average amount of the dilute serum was 0.181 c.cm. Fifteen patients, convalescent after a moderate illness, gave an average of 0.276 c.cm. Two patients seriously ill with temperatures of 104.3° F. and 105° F. required 0.4 c.cm. to bring about complete hemolysis. We would not like to draw any very definite conclusions where we are dealing with such small fractional differences. This lessening of complement has been noted in other infectious diseases and may be important in the questions of immunity in influenza. Dr. Frost carried out a number of cutaneous tests after the method of Von Pirquet, using a polyvalent, weakly alkaline emulsion of influenza bacilli in 25 per cent. glycerin. Eleven convalescents were tested and none of them showed any local or general reaction. The suggestion that these results may indicate an increase in resistance is discussed in another place. A number of strains of pneumococci which we had isolated from our autopsy cases were differentiated by the agglutination method. Type I was found 3 times; type II, 10 times; type IV, 9 times. Four showed agglutination with both type I and type II sera. Type IV pneumococcus was isolated in one case from the right and left bronchus as well as the lung. In another case the same type pneumococcus was recovered from the lung and pleural fluid. These results are similar to those found by numerous workers.

The Hemophilic Bacteria

The discovery by Pfeiffer of the hemophilic character of the bacillus found by him in cases of influenza opened up a new group of micro-organisms known as the hemophilic bacteria. Davis (1915) has laid particular stress on the group character of these bacilli, and the more they are studied the more clear does it become that there are several distinct members. The B. influenzæ is by far the most important as well as the most frequently found of the group and is considered as the type organism.

All these bacteria require for their growth the presence of some form of hemoglobin. The actual amount necessary may be very small, and Davis suggested that it may have a catalytic action. A great deal of work has been done in attempts to discover just what portions of the hemoglobin are necessary to bring about this phenomenon. In our discussion on media for the influenza bacillus we will briefly describe some of the various hemoglobin preparations that have been used successfully. It must at this point be emphasized that blood is very useful in many media to stimulate the growth of a great variety of bacteria, and the transfers made from such luxuriantly growing cultures may grow very poorly or not at all on ordinary media, and this might easily lead to erroneous conclusions on the hemophilic character of the organisms studied. There are certain bacteria which grow so much better on media containing blood that such media are sometimes necessary for their isolation, although after a few transfers they will grow on ordinary media. This is true for bacillus pertussis, and throughout the literature a good deal of confusion has arisen in not recognizing this temporary hemophilic character of certain bacteria. The true hemophilic bacteria do not grow except in the presence of hemoglobin in some form or other. The problem becomes almost academic when we consider the small amounts of hemoglobin that are necessary. Davis has shown that a dilution of 1 in 180,000 is sufficient, and in the interesting discussion between Cantani and Ghon and Preyss it was demonstrated that hematin or other hemoglobin product was necessary in the agar before B. influenzæ would grow in the presence of other bacteria, and that this hematin could be derived from the blood in the meat which was used in making the basic infusion.

Symbiosis.—The fact that other bacteria can bring to growth the influenza bacillus on media otherwise unsuited to its needs brings up the interesting problem of symbiosis, which is one of the most important characters of the influenza bacillus. Not only do other bacteria make possible the growth of B. influenzæ on media on which the influenza bacillus will not grow, but they stimulate a better growth on blood agar and other more or less favorable media. Grassberger first noted this stimulating character of other bacteria and described and illustrated the very large colonies of B. influenzæ which develop in the neighborhood of colonies of staphylococcus and other bacteria. Staphylococci killed by heat were found to have a similar effect. Meunier nicely described this phenomenon by using the term satellites for the circles of B. influenzæ colonies which develop about the colonies of other bacteria. A great number of workers have since noted this characteristic relationship between B. influenzæ and other bacteria, and occasionally have laid stress on its importance in the problems of the infections by the influenza bacillus. Allen particularly emphasized the probable importance of this in discussing the problem of carriers of B. influenzæ as sources of danger. There seems no doubt that this symbiotic relationship depends on so altering the hemoglobin products as to render them more readily available for the influenza bacillus. This is indicated by the fact that on various media containing hemoglobin, altered so that it encourages the growth of B. influenzæ, no such symbiotic stimulation can be demonstrated. This phenomenon is quite peculiar to this bacillus, distinguishes it from most of the other members of the group, and should be always determined before an organism is classed as B. influenzæ.

Other Hemophilic Bacteria.—The question of a pseudo-influenza bacillus was first raised by Pfeiffer and has been studied by many workers after him. Grassberger, who carefully investigated this problem, worked more particularly with two strains showing the extreme of variation between the small characteristic morphology of the B. influenzæ and the thread forms supposed to be characteristic of the so-called pseudo-influenza bacillus. The great majority of workers have agreed with him in concluding that this morphological variation is not sufficient nor constant enough to justify separating two such groups. Nevertheless many reports indicate peculiar tendencies of certain strains toward thread formation. There seems to be suggestive evidence that the organism described by Cohen in 1909 under the name B. meningitidis cerebrospinales septicemicus is different from true B. influenzæ. Although the cultural characters were apparently identical, this organism was definitely pathogenic for guinea pigs and rabbits. The involvement of joints in the cases reported by Longo and others would suggest a greater pathogenic power for these strains. Prasek and Zatelli reported a similar bacillus from meningitis, and Davis found that his meningitis strains were more pathogenic for rabbits than were others. Wollstein has studied this question very carefully and found a marked difference between the strains from the meninges and those from the respiratory tract in their pathogenicity for rabbits. The strains with a tendency to thread formation were usually also those grown from the meninges, but she concluded from the results of serological tests that all strains of B. influenzæ are of one race, irrespective of their origin or virulence. The question is still an open one, as Batten and others described strains from the meninges which are non-pathogenic, and Ritchie found his strains from meningitis pathogenic for guinea pigs but not for rabbits. The irregularity and wide divergence in the results of blood cultures may have a definite relationship to these differences in the pathogenicity of strains.

Other hemophilic bacteria include the bacillus described by Friedberger under the name of B. hemoglobinophilus canis. This organism is to be found in the preputial secretion of dogs. It does not show the phenomenon of symbiosis, and I have found that it grows rather more freely and is more resistant to drying than is the influenza bacillus. Krage has confirmed Friedberger’s findings growing this bacillus from 60 per cent. of his dogs, and believed it a pyogenic organism just as B. influenzæ may be.

The hemophilic and hemolytic organisms described by Davis, which he isolated from pathological urine, were non-symbiotic and non-pathogenic. Koch has described a similar organism from puerperal infection. Whether the hemophilic organism described by Thalhimer from the uterus in a case of puerperal infection, those found by Cohen in urethral discharge in one case and the pelvic exudate of another, and the findings of Kretz in pyelitis, Wright in pyelonephrosis and Klieneberger in cystitis cases, possibly refer to this same bacillus is, of course, uncertain. Pritchett and Stillman found a somewhat similar bacillus, which they called Bacillus X, from the mouths of 24 persons. It was hemophilic and hemolytic, stouter than B. influenzæ and showed long tangled threads in blood broth. It was non-pathogenic and is probably the same as Davis’ organism.

Davis described another hemophilic bacillus from a patient with purulent foci which was non-hemolytic and non-symbiotic. It was grown from an abscess of the shoulder joint, the blood and the bronchial secretion of an infant. Cyanosis was a marked feature of this case. Paranhos described a hemophilic bacillus from meningitis, which, however, was Gram positive, and Moon reported an anærobic hemophilic bacillus from an infection of the ethmoid sinus. The work of Jordan would suggest that there may be two groups of B. influenzæ based on the indol production.

Morphology.—The morphology of B. influenzæ has received more than usual attention. In what we consider its characteristic form, it is an extremely small bacillus, usually single but sometimes in pairs, and not infrequently exhibiting polar staining. In direct smears, where there are many bacteria present, they are frequently arranged in the schools so frequently described. The development of thread forms is today considered quite characteristic for B. influenzæ. The organisms vary from moderately long bacillary forms to very long twisted or curled threads suggesting leptothrix. In such cultures chains of tiny bacilli are also quite often noted. At the other extreme we have exceedingly tiny coccoid forms, resembling in size the B. bronchisepticus, which, as Ferry has shown, are small enough to pass through many grades of filters.

It is the thread forms, as discussed above, that have received most attention in relation to the so-called pseudo-influenza bacillus. The observations of Wollstein, Lacy and many others showed these forms to be common in meningeal infections and that, as a rule, they are more pathogenic for animals than other strains. Another interesting and important observation is that emphasized by Dick and Murray of the possible confusion of these forms with Gram negative leptothrix. That this confusion is liable to occur is illustrated by reports such as Macdonald finding leptothrix in a meningeal infection, now looked upon as an example of influenzal meningitis, and the probable B. influenzæ reported by Dick, and, as quoted by Dick and Murray, the finding of a Gram negative leptothrix as the cause of broncho-pneumonia by Kato. The 2 per cent. leptothrix reported by Nuzum and his co-workers from the recent epidemic may be still another example. Equally important is the recognition of the great frequency of this thread development in the majority of B. influenzæ cultures on ordinary blood agar media, or even in the water of condensation of fresh blood agar tubes. The delayed growth of this bacillus on ordinary blood agar would lead to its being frequently overlooked unless smears are made, and the irregular thread forms are recognized as being the B. influenzæ. This development of thread forms was particularly noted in my work before pickings were made to the Voges heated blood agar, but because I had been forewarned by discussing these morphological variations with Lacy, I was able to recognize them as forms of B. influenzæ. Most of my early isolations showed these predominating, and they were also noticed in cultures sent from the Public Health Laboratory at Washington. These cultures on further transfer, however, showed in 24 hours the typical small form on ordinary blood agar as well as on the Voges medium. On the latter the development of thread forms was greatly delayed and frequently did not appear at all, although after long periods other abnormal, swollen and irregular shapes sometimes developed.

Media in Growth of B. Influenzæ

The discovery of the hemophilic character of B. influenzæ has been confirmed by a long list of investigators. The agar smeared with pigeon blood as used by Pfeiffer has not, however, been found fully satisfactory and many modifications have been made. The fact that hemoglobin in some form is necessary for the growth of these bacteria has led to a great deal of study in attempts to discover the chemical part, or parts, essential for this purpose. Hemoglobin in very small amount, as shown by Davis and others, is sufficient to make media suitable for growing B. influenzæ. This fact has led to much confusion, owing to the difficulty of eliminating all possible sources from which some form of hemoglobin might enter the media. Kitasato used a glycerin agar and succeeded in growing the influenza bacillus for 10 transfers. Pielicke, however, did not consider that Kitasato was actually dealing with the influenza bacillus, but that he as well as Babes, Bruschettini and Markel had most probably streptococci in their cultures. Besson held the same view of Kitasato’s organism. It would further appear from the illustrations of Klein that he also grew streptococci and not the B. influenzæ. The first culture of the influenza bacillus was probably obtained by Bujiwid in February, 1890. He grew on agar smeared with the spleen pulp of an influenzal patient a tiny bacillus which he was unable to grow on blood free medium, but he did not appreciate its importance until Pfeiffer’s article appeared. Teissier in his book on “L’ Influenza en Russie” mentioned this culture.

The hemophilic character of these bacteria indicates that they are rather strict parasites, and despite the researches of Nastjukoff with various egg media, and Cantani with a number of supposedly non-hemoglobin additions to the agar, as well as the studies on symbiosis, with other bacteria, by Cantani, Neisser, Luerssen and many others, it still remains true that some form of hemoglobin is necessary for their growth. Fresh blood either incorporated in the medium or smeared on the surface is not the best medium for these bacteria. Altered hemoglobin is much more favorable, and a variety of methods have been devised to bring about those alterations which stimulate the growth of B. influenzæ. One of the earliest, as well as one of the very best, of these is the method of Voges, who added blood to melted agar at a temperature of about 100° C. I have found this medium exceptionally suited to growing B. influenzæ, and I consider it excellent for the primary culture from the original material, for pickings from plates and to obtain a heavy growth of B. influenzæ for any purpose. The medium was used by Delius and Kolle (1897), Grassberger (1898), who spoke very highly of it, and Paltauf (1899), who said that the use of this medium made the demonstration of B. influenzæ possible when only a very few were present. A great many other workers have used it with success, and during the recent epidemic it has gradually found its place. Levinthal’s medium (1918) is practically the same, although he boiled and filtered the agar after the addition of the blood. The growth of B. influenzæ on the Voges agar can properly be described as luxuriant, and to anyone only accustomed to the use of ordinary blood agar it is an agreeable surprise to see this supposedly delicate bacillus growing so remarkably well.

Various other methods have been used to bring about this beneficial change in hemoglobin. Gioelli (1896) used a medium made up of 1.1 per cent. hemoglobin and 21.5 per cent. malt extract. This is reddish brown in color, becomes clear when neutralized with potassium hydrate and remains so on heating. This added to agar is reported as very favorable in growing this bacillus. Ghon and Preyss described a medium made up of meat, peptone, salt and agar prepared in the ordinary way, but not filtered for at least a week, and then only roughly. This medium is favorable for symbiotic growths. He further used beef blood heated in a soda solution and blood heated in water as hemoglobin preparations to be added to agar. Thalhimer found an amorphous hemoglobin medium to be more favorable than when a purer hemoglobin was used. W. F. Robertson found a hemoglobin agar, prepared by allowing sheep’s blood to clot, decanting off most of the serum, freezing and then thawing what remains and adding 1 c.c. of this to an agar tube at about 60° C., to be very favorable for the growth of B. influenzæ. Cantani used a blood treated with pepsin and hydrochloric acid, digested some days in the incubator, filtered and made weakly alkaline. This mixture was heated for a few minutes, refiltered and added to the medium. He speaks of it as extraordinarily good for B. influenzæ. Blood treated with trypsin has been used by Matthews, Averill, Young and Griffiths, Harris, A. Fleming and others. Fleming further found that this alteration in hemoglobin can be brought about in a number of other ways. Blood boiled in agar (suggesting the Voges agar) and the tubes slanted while hot, blood boiled in water, the clotted blood precipitated and the clear fluid added to agar, or more rapidly by adding equal quantities of sulphuric acid to the blood and a similar amount of potassium hydrate he obtained altered blood suitable for media. He reported that by any of these methods he could obtain a medium very stimulating to the growth of B. influenzæ. By the addition of brilliant green (1 in 500,000) he inhibited the growth of staphylococcus, streptococcus and pneumococcus. For storing cultures of B. influenzæ Fleming found a minced meat medium with the addition of blood to be the best. I have found this medium without the blood to be an excellent one for keeping a great variety of cultures. Bernstein and Loewe have reported the use of gentian violet (1 in 5,000) for the same purpose as the brilliant green used by Fleming. Avery’s oleate blood agar medium he reported to be largely selective. It checked the growth of pneumococci and streptococci, but gave luxuriant growths of B. influenzæ. Pritchett and Stillman have used it with excellent results recovering B. influenzæ from a very high percentage of the cases studied.

The use of symbiotic bacteria has been extensively studied in investigations of the biology of B. influenzæ, and it has been shown, as noted elsewhere, that such accessory bacteria will bring to growth B. influenzæ on media otherwise quite unsuited to its needs. It has been further found that on various preparations of hematin agar, on which B. influenzæ refused to grow, such media could be rendered favorable for their growth by the addition of living or freshly killed cultures of staphylococcus and many other bacteria. And although the method is well known, it has not been extensively used for the purposes of isolation. Many of the workers, however, have pointed out the importance of looking for growth of the influenza bacillus in the neighborhood of the more easily grown bacteria which almost always develop in cultures from the respiratory tract. Grassberger has particularly studied this problem and has made practical application of the method. Accidental contamination of plates with air bacteria have made possible, in some instances, the isolation of B. influenzæ—as, for example, in the finding of Heyrovsky from a case of empyema of the gall bladder—while other workers have pointed out the difficulty of demonstrating growth where B. influenzæ is pure in the material cultured, and the comparative ease and relative luxuriance of growth where other bacteria are present. To just what this stimulating effect is due has been much discussed, and it is generally agreed that the hemoglobin is markedly changed and rendered more available by the action of these germs. It is to be noted that on a medium containing blood altered by heating or by the various methods as described by Fleming the foreign bacteria no longer show any symbiotic action on B. influenzæ. Grassberger considered the effect of the bacteria on the blood to be the same as that of heating. Allen laid particular stress on this symbiotic character. He used a staphylococcus, either living or killed, in making his cultures and noted the difficulty of growing B. influenzæ from material in which it occurred pure. W. F. Robertson made use of these facts of symbiosis for both isolation and stimulation of growth. He employed alternate drills of M. catarrhalis or pneumococcus with the B. influenzæ, and Brown and Orcutt used strains of hemolytic streptococci for the same purpose. The latter authors considered that the beneficial effect of the streptococci was merely due to the setting free of the hemoglobin. The fact that similar results are to be obtained by the use of non-hemolytic bacteria as well as forms giving green color changes to the blood makes this explanation untenable. In my own studies I have confirmed the results of several previous workers. I have found that B. influenzæ is stimulated in its growth by the presence near it of colonies of staphylococcus pyogenes aureus and albus, pneumococci, streptococcus viridans and hemolyticus and other bacteria. The largest colonies of the bacillus I have obtained were those growing near the periphery of a colony of an air nocardia. I have also noted that emulsions of a staphylococcus killed by boiling for five minutes, when added to ordinary blood agar, had a marked stimulating effect, although no evidence of hemolysis was present. This effect was practically absent if the emulsion was boiled for 15 minutes, or after being killed was left at room temperature for several days. There was no evidence of these stimulating effects by any of these methods when heated blood agar was used, the colonies on this medium growing equally large by themselves. Comparative studies of the effect of different bacteria can be simply carried out as follows: Smear evenly the surface of an ordinary blood agar plate with an emulsion of B. influenzæ. Seed this plate at various points with minimal amounts of the various bacteria. After various periods of incubation the size of the B. influenzæ colonies about the other bacterial growths can be estimated, and impression preparations on cover glasses will give very interesting pictures.

The growth of B. influenzæ in primary cultures from sputa and similar sources is to be explained by the probable presence of traces of blood or altered hemoglobin as well as the symbiotic relationship with other bacteria. Fichtner used fresh heated sputum (60 to 65° C.) in place of blood, and Richter a medium made with sterilized pus. Parker, in her study of a filterable poison produced by the B. influenzæ, found veal infusion broth with 10 per cent. defibrinated blood heated to 75° C. until the blood coagulated and settled on standing to be the best for the purpose. Jordan in his study of indol production by these bacteria used a meat infusion broth with 5 per cent. sheep’s blood added at 90° C. or over and filtered through cotton or paper. Wittingham and Sims noted that in using blood from influenza cases the bacteria frequently did not grow, more especially B. influenzæ; and Rivers found human blood poorer than cat or rabbit blood for growing this organism, as did Minaker and Irvine. It would seem clear from this review of some of the suggestive work on the methods of growing B. influenzæ that little attention should be given to the results of many workers, where ordinary media were used, particularly when the difficulties of isolation were not appreciated.

B. Influenzæ as a Pathogenic Bacterium

If B. influenzæ is the causative agent in clinical influenza, there is certainly ample evidence that it is pathogenic to man. The symptoms of toxemia, which are so manifest in the pandemic disease as well as in the sporadic cases, would indicate that the etiological agent is markedly toxicogenic. Animal experiments by Pfeiffer, and a long list of investigators following him, would seem to show that the majority of cultures of B. influenzæ do not have any power of establishing themselves in the animal tissue. Killed cultures showed equally as high toxic effects as the living, and so it was generally concluded that many of the general effects in influenzal infections were of a toxic nature.

There are many exceptions to the above-mentioned failures to produce infections in animals. Cantani obtained very constant positive results by subdural injections. He first clearly showed that killed cultures were markedly toxic and that virulence could be raised very definitely by animal passage. By injecting brain emulsion with a culture he obtained a subcutaneous abscess in a rabbit which after eight days still contained the living organism. Nastjukoff found that animals with a lowered resistance, or definitely ill from, for example, an artificial tuberculosis, became infected while others did not. Jacobson showed that B. influenzæ injected with streptococci caused a definite mixed infection, and that after six passages the influenza bacillus alone could produce a fatal infection. Saathoff (1907) confirmed Jacobson’s findings and found pneumococci equally effective. Davis (1915) also confirmed the principle established by Jacobson of the symbiotic relation of other bacteria to infection with B. influenzæ. He used a culture of a non-virulent staphylococcus pyogenes aureus, and was able to produce death invariably in guinea pigs after intraperitoneal injection. From the heart’s blood, as a rule, only the hemophilic bacillus was recovered. He also found animal passage increased the virulence, and further that M. catarrhalis and an avirulent streptococcus had the same effect as the staphylococcus. Slatineanu (1901) found that he could infect animals with B. influenzæ if the cultures were injected along with weak solutions of lactic acid, and that after animal passage by this method the bacillus became more virulent and would eventually kill by itself. It must not be forgotten in this connection that strains of B. influenzæ from meningitis cases are frequently definitely pathogenic for animals. The importance of considering these various factors in a discussion of infection by this organism is, of course, very evident. Ecker found his strains pathogenic for mice after subcutaneous injection, and the bacilli were readily obtained from the heart’s blood. Spooner and his co-workers from their results of more than a hundred intraperitoneal injections concluded that the organism is not pathogenic for mice.

In all animal experiments it is of the greatest importance that the bacteria be known which may interfere in the experiments through spontaneous infection (often liable to be induced by the injection) from the animal’s own flora, as well as the greater susceptibility of previously diseased animals (Nastjukoff). It would appear from the results of Bruschettini and Cornil and Chantemesse in the early days of the influenza bacillus, and those of Lamb and Brannin in their recent study, that these authors did not seriously consider the spontaneous infection of guinea pigs and rabbits with B. bronchisepticus or the bacillus of rabbit septicæmia, both morphologically, very similar to B. influenzæ. Rosenow in his experiments with streptococci from cases of influenza has also apparently failed to realize the importance of the lung lesions produced by the B. bronchisepticus in guinea pigs as reported by Theobald Smith, myself and many others.

Parker has found a filterable poison from the influenza bacillus which developed rapidly (6 to 8 hours) in a special heated blood broth medium, deteriorated rapidly even in the cold, and killed rabbits in quantities of 2 c.c. in from 1 to 3 hours. Rabbits could further be immunized against this poison, and their sera protected other rabbits against fatal doses. This is the first time that a true powerful toxine has been obtained. Couret and Herbert obtained toxine from B. influenzæ in Avery’s oleate broth. Huntoon and Ross also clearly demonstrated toxine production by this organism so that it would appear, with this confirmation, that the B. influenzæ can be definitely classed among the toxine producers. Toxemia being the most striking clinical characteristic of influenza, we have in these findings very strong evidence of the etiological importance of this hemophilic bacillus to the disease. A very interesting observation was made by Latapie that the serum of a goat immunized against influenza bacillus is toxic if it is used shortly after the injection of the microbes, but that this toxicity is absent three weeks after the last injection. It would appear to me that the evidence of a filterable virus from the secretions of the respiratory tract does not eliminate the very probable toxine from such materials. The production of toxine by this organism probably depends, as is the case with very many of our toxine formers, on the most favorable combinations of conditions. That it is not readily formed in artificial cultures, or that it is very unstable if formed, is evidenced by the frequent failures of a great many workers. It has been suggested that different symbiotic conditions in the respiratory tract determine the amount of toxine produced. Huntoon found a high toxine production in mixed cultures with streptococci. This, however, does not appear to be necessary, as there is ample evidence of severe toxemia from pure infections with B. influenzæ in various parts, such as the accessory sinuses of the head, the meninges, the lungs and other parts of the respiratory tract.

It is not fundamentally necessary that a toxine producing organism be present in overwhelming numbers before it can be accepted as the cause of the toxemia. Nor, on the other hand, must we have toxemia every time the organism is found. The prevalent idea among bacteriologists would appear to be the reverse of what I have just stated. It would, indeed, be extremely difficult to make bacteriological diagnoses of a great many of our diseases, where the etiological factor is well established, if these conditions were required. We do not do so, for example, in diphtheria, examinations of stools for typhoid, nor in infections with the tetanus bacillus. We recognize carrier cases of meningococcus, B. typhosus, hemolytic streptococci and many others, without detracting seriously from their importance in definite types of infection. Formerly the specificity of the different bacteria for definite disease processes was very rigid, but today we interpret more broadly the finding of gonococcus in endocarditis, the meningococcus in bacteremia, B. typhosus in osteomyelitis, streptococci and pneumococci in all manner of infections and many other bacteriological results. True it is that the various bacteria show predilections for attacking certain tissues, but the varying susceptibilities bring about the greatest variations in the manifestations of these infections.

The B. influenzæ is not confined to the causation of severe pandemic or epidemic influenza, but includes in its field purulent bronchitis, meningitis, sinusitis, conjunctivitis and many other pathological processes. It further should be recognized as a relatively frequent cause of complications in measles and other diseases.

Infections of the Respiratory Tract

The disease influenza is primarily an infection of the respiratory tract. It varies from one of the most acute and fatal diseases we know of through all grades of severity—from chronic infections lasting over years to the familiar three or five day fever. This graduation is to be found more or less marked in all our bacterial infections, but would seem to be not generally recognized or appreciated as occurring in infections with the influenza bacillus. That Pfeiffer was dealing with one phase of the disease when the influenza bacillus was discovered does not invalidate the results of numerous workers which have been added since then.

Probably the greatest confusion in attempts to get a clear picture of this protean disease has been and is a non-recognition of influenza as a frequent complication of other diseases, such as measles (Jochmann, Susswein, Tedesko and very many others). The second cause for this confusion has been the misinterpretation of the facts demonstrating the rather frequent occurrence of carriers. During an epidemic the vast majority of patients show the disease as an upper respiratory infection of varying degrees of intensity, but which usually subsides after periods of from three to five days of fever. Along with this we have other graded manifestations of further involvement of the tract with laryngitis, bronchitis, bronchiolitis and all degrees of broncho-pneumonia. To prevent the severe lung involvement prompt treatment must be carried out, under which rest in bed is by long odds the most important. This will be discussed in another paper of this series, and was particularly well demonstrated in the results at the Naval Hospital as verbally reported to me by D. G. Richey. The interesting point is that the infection can be controlled, but this does not indicate the etiological factor as different from that acting in the more severe cases.

The epidemiological evidence would seem to show very clearly that the incubation period is approximately two days, and that a period of six weeks is the usual limit for the severe wave of the epidemic in different localities. In my opinion, during this period every exposed individual in a community has received the influenza bacillus in the respiratory tract, and that all the susceptible individuals are attacked and show more or less evidence of the infection. As a consequence of this general distribution we have great numbers of individuals carrying the organism, and the aftermath is to be noted in other and later manifestations of the same infection.

Sporadic cases of influenza appear during inter-epidemic periods and more or less healthy carriers are frequent. Scheller’s study in Königsberg showed, if we can rely on his figures, that the carriers were very numerous during an epidemic year (winter 1906-1907), being 24 to 33 per cent.; that as the epidemic became less widespread (winter 1907-1908) it fell to 10 to 13 per cent.; as it was disappearing (summer 1908) he found only 1.5 to 3.3 per cent.; while when the epidemic was completely over (winter 1908-1909) there were no carriers of B. influenzæ found. These results are taken from studies of sputa and throat smears of 138, 218, 155 and 185 cases, respectively, for the periods mentioned. The monumental work of Tedesko, who reported the results of 1,479 cultures, covering 11 years (1896-1906), would indicate that B. influenzæ is continually present in the population. However, in carefully analyzing his results, it is very clear that in the great majority of his cases it was of definite etiological significance. Lobular pneumonia, acute, purulent and chronic bronchitis, and most frequently clinical influenza, are the prominent diagnoses in all his tables. He was able to grow B. influenzæ repeatedly from individual patients for many months.

Lord in similar studies (1902, 1905, 1908) brought out somewhat similar facts. He laid particular stress on the cases of chronic bronchitis with numerous B. influenzæ in the sputum and a probable confusion of these with pulmonary tuberculosis. He was able to follow a number of his patients for several years. B. influenzæ was grown in culture from the sputum of one of these in 1902; in November, 1903; in February, 1904, and in February, 1905. In other cases the organism was shown to be present by culture practically continuously for months and even years. Lord, with Scott and Nye, in a recently published article (1919) reviewed his former results and showed a relatively high incidence of B. influenzæ in the respiratory tract of apparently healthy people. Davis studied 534 cases, further indicating the prevalence of this organism in the community.

The B. influenzæ has been recovered from the respiratory tract during the clinically pure influenza, from the sputum and lung in influenzal pneumonia, and from the purulent sputum in all grades of bronchitis. These should all be looked upon as true infections by the influenza bacillus, the varying manifestations merely differing with the resistance of the individual. In the epidemic in the fall of 1918 pneumonia was the outstanding feature. Preceding this in the English publications we have reports of outbreaks of purulent bronchitis. Macdonald and his co-workers, finding the B. influenzæ frequently present, considered the condition as one indication of a virulent infection by this organism. Hammond, Rolland and Shore reported similar cases, and Abrahams and his co-workers looked upon the cases of purulent bronchitis as occupying a position, without any definite line of demarcation, between those with definite broncho-pneumonia on the one side and those with simple bronchial catarrh on the other. H. E. Robertson emphasized the serious nature of influenzal purulent bronchitis and the almost epidemic character and rather high mortality of the outbreak in the winter and spring of 1917-1918. There were also numerous mild outbreaks of influenza before the overwhelming culmination of the last three months of 1918, as reported by Orticoni and many others and noted by Johnston in this series of papers. Greenwood in an epidemiological study emphasized the point, previously made evident by Parsons for the pandemic of 1889-1892, that the mass attack is preceded by numbers of individual cases. In this country it was noted during the winter of 1917-1918 and the following spring that the B. influenzæ was rather frequently found in the respiratory infection in our army camps (Soper, Cole and MacCallum and others).

It is well recognized that when the actual epidemic struck there were comparatively few bacteriologists familiar with the B. influenzæ. The real difficulties of isolation, the more favorable media, the facts of symbiosis, the importance of carriers, the varying manifestations of the infection and many of the other vitally important points, although more or less fully reported in the literature, were nevertheless practically unknown. It was my own experience, and that of many others. This must be seriously considered in analyzing many of the reports on bacteriological findings throughout the period of the severe wave and even after.

Results of Others During the Recent Pandemic

It will be impossible to review the numerous reports on the recent epidemic that have appeared. Many of these can be discounted, as far as the finding of B. influenzæ is concerned, for the reasons mentioned above. The often quoted report of Little, Garofalo and Williams, who did not even use a hemoglobin medium, will serve as an example. Little attention should be given to others where the large numbers of cases precluded the requisite time and media necessary for such a difficult problem. Friedlander and his co-workers in their report from Camp Sherman made no mention of the number of sputa, throat swabs or autopsies which they examined bacteriologically. The incidence of influenza showed a total of 10,979 cases, 2,001 of pulmonary œdema or pneumonia and 842 deaths. They recorded one culture from the sputum with pneumococcus predominating which gave two colonies of B. influenzæ, and this bacillus was grown from the lung exudate at one autopsy. Their conclusions that “B. influenzæ (Pfeiffer) has not been demonstrated as the causative organism” is certainly true from their results, but that “the frequency of its detection has not exceeded the frequency of its existence under normal conditions” can hardly be considered as established, if we accept the many results mentioned above as indicating its presence during inter-epidemic times, unless they mean by normal conditions practically complete freedom from this organism.

The prevalence of B. influenzæ in various sections of this country may be indicated by the following reports chosen from many available ones. Keegan, from the First Naval District Hospital, found B. influenzæ 19 times from 23 in cultures grown from the lungs. In 6 cases these cultures were pure. Medalia reported from Camp McArthur the following. Out of 2,279 sputa of influenza suspects, 76.8 per cent. showed “B. influenzæ” in smears, and 445 sputa from cases of broncho-pneumonia showed it in 54 per cent. It was found in culture in only 10.6 per cent. of these last cases. He considered sputum smears of practical diagnostic help. He further grew B. influenzæ twice from the blood during life, once with a pneumococcus and once alone. Necropsy cultures gave B. influenzæ in 2 of 3 cultures from the brain, 19 of 34 from the heart, 19 of 36 from the spleen, 54 of 65 from both lungs, 50 of 62 from the right pleura and 47 of 62 from the left pleura. The percentage of positive results ranged from 53 in the spleen to 83 in the lungs. Nuzum and his associates only found B. influenzæ in 4 of 100 cases from the bronchial secretions, but it is interesting to note that he grew it in practically pure culture from both lungs of one case at autopsy. Synnott and Clark in Camp Dix found streptococci and pneumococci predominating, and, although making no particular effort to study the B. influenzæ or determine its frequency, they found it in the majority of cases when it was looked for. Blanton and Irons reported as follows from Camp Custer. From cultures of the nose and throat of 357 examined before the epidemic struck, B. influenzæ was found in 5.1 per cent.; in 366 throat cultures of influenza cases without physical signs of pneumonia the same organism was grown in 44, or 8 per cent.; sputa typed for pneumococci 740 times from influenza cases with pneumonia gave isolations of B. influenzæ 38 times, or 5 per cent.—8 times alone, but here it should be remarked that these latter isolations were only attempted after the organism was suspected from the morphological picture of the smears; from 280 autopsies B. influenzæ was recovered 8 times from the lung and 3 times from the heart’s blood. This report covered the period from the outbreak of the epidemic, October 5 (or as given by Soper, September 30) to October 22, at the outside a period of 22 days. During this time 366 throat cultures, 510 blood cultures, 740 sputa typed for pneumococci, 280 autopsies with cultures from both lung and heart’s blood, made a total of primary cultures of well over 2,000. The technical difficulties would make it almost impossible to handle such a mass of material and get reliable results for the incidence of B. influenzæ.

Brem, Bolling and Casper in Camp Fremont found B. influenzæ in 259 from 537 selected cases in swabs from the nasopharynx. It was also noted in a fair number of other examinations. Opie and his co-workers found B. influenzæ to be very frequent at Camp Pike. Spooner, Scott and Heath isolated B. influenzæ at Camp Devens from the sputa of 104 cases, from nasopharyngeal swabs in 11 out of 18 attempts and from the pleural fluid 8 times out of 45, twice pure. From 37 autopsies they found B. influenzæ in 23 and in pure culture in at least 1 lobe of the lung in 16. From 82 blood cultures at autopsy B. influenzæ was recovered twice. Nichols and Stimmel studied lung punctures during life and grew the B. influenzæ from 7 out of 10 attempts, 5 times in pure culture. Stone and Swift at Fort Riley found B. influenzæ in 18.7 per cent. of 928 sputa and in 5.2 per cent. of 77 sputa from fatal cases. He recovered it from autopsy material; 21 times from 51 lungs, once alone; twice from 26 pleural fluids; twice from 30 heart bloods; 19 times from the sinuses of 40, and 9 times from the ear and mastoid of 17 cases.

Lamb and Brannin at Camp Cody examined 80 typical cases early in the epidemic. They found B. influenzæ predominated in 46 per cent. being present with pneumococci on 41 per cent. of the plates. They also grew the influenza bacillus from a fair number of other cases.

Wollstein and Goldbloom in the Babies Hospital of the City of New York found the B. influenzæ in 13 of 17 sputa during life and in both lungs of all 18 autopsies as well as in the heart’s blood of one. Kotz found it in half of his 30 cases. Pritchett and Stillman grew the influenza bacillus from 41 of 49 cases of influenza, from 40 of 43 cases of influenza with broncho-pneumonia, from all of six other broncho-pneumonia cases and from 11 of 20 cases of lobar pneumonia, making a total of 98 positive findings from 118 or 82 per cent. They further found 25 positives from 54 convalescent and 74 from 177 normal sputa. Wolbach found this organism in pure culture in one or more lobes of the lungs of 9 from 23 cultured cases. It was demonstrated in 23 of 28 either by culture or in section.

Similar results are to be found in reports from Great Britain. Martin noted a great increase in the numbers present as the sputum became more purulent. Hicks and Gray found B. influenzæ by culture in 75 per cent. of their cases. They were seen in direct smears in only 70 per cent. Gotch and Wittingham considered M. catarrhalis to be the etiological factor as it was found in all of their 50 cases. B. influenzæ was grown in 8 per cent., although B. influenzæ-like bacilli, were seen in 62 per cent. of their smears. Averill, Young and Griffiths studied the sputum from 41 cases and found B. influenzæ in 32. It is interesting that Macdonald and Lyth determined the incubation period to be 41 hours as a minimum in their own experience and that from the posterior nares of one of them B. influenzæ was obtained.

Schofield and Cynn found the B. influenzæ in Korea. Kraus in Brazil found it in the sputum in 62 per cent. of his cases of influenza. It was also found in the organs of 27 who had died, being in pure culture in five. It has further been found in France, Italy and practically all parts of the world where investigations have been made. The German literature is at present only available in the report of the British Medical Research Committee which is written in a more or less popular manner with a rather strong tendency against the importance of B. influenzæ. Dietrich, Simmonds, Bergmann and others, however, found B. influenzæ rather frequently. Such quotations as “Uhlenhuth, a diehard of bacteriologic orthodoxy, has clearly shown signs of uneasiness” and “one empyema and one throat swab yielded the looked for growth” will indicate why this review is of little use. It is certainly necessary to “look for” the B. influenzæ to get results of any worth.

Secondary, ancillary or symbiotic bacteria are of cardinal importance in these infections. It has been considered by some writers as characteristic for the influenza bacillus to be followed so frequently with such a variety of secondary invaders. Sahli looked upon the complex of B. influenzæ, pneumococcus and streptococcus as the true etiological cause of influenza. Abrahams and his associates discussed the symbiotic effect of the B. influenzæ in raising the virulence of pneumococci previously present in the patient and many other investigators lay stress on these symbiotic relationships.

Pneumococci appear to be the commonest of these secondary micro-organisms judging from the various published reports, but the fact must not be overlooked that, particularly in America, the typing of pneumococci has drawn a disproportionate attention to this group. Hemolytic streptococci have received much attention (Ely and his co-workers and several others). M. catarrhalis (Gotch and Wittingham and several of the British writers), members of the B. mucosus capsulatus group (Nichols and Stimmel, Rucker and Wenner), staphylococcus aureus (Patrick), various ill-defined streptococci (Rosenow and several British writers), capsulated cocci apparently different from pneumococci, B. pestislike forms and many others have been given more or less attention, often as clearly recognized secondary infections, but not infrequently as of primary significance.

B. influenzæ, however, is the organism most regularly found in this pandemic where carefully looked for, and the evidence of its lowering the general resistance to bacterial invasion is very strong. The experiments of Ghedini and Fedeli showing the effect of the toxine on muscular tone and those of Ghedini and Breccia who found a similar effect on blood vessels are worthy of note.

The fact that the flora differs so widely in various regions is what one might expect and many investigators have emphasized the significance of this. Bacteria in the mouth and throat are readily transmitted from individual to individual and under the conditions in the training camps and our modern life, the development of local flora is not surprising. That it is of very great importance is recognized by all and it is often a determining factor in the severity of the infection. Nevertheless, influenza in this pandemic has been almost equally severe whatever the secondary organism may have been.

I have discussed in another place the suggestion of the stimulating effect of various bacteria on the growth and toxine production of B. influenzæ. Huntoon showed the effect of hemolytic streptococci in cultures to be helpful in toxine production. An important point, however, is that no one bacterium has been shown to be exclusive in thus affecting the growth on media of the influenza bacillus, and in the animal experiments in raising the invasive and pathogenic power of this organism the same appears to be true. The infection in influenza, in the vast majority of cases, rapidly becomes a mixed one. The secondary organisms at times completely dominating the field, at least as far as numbers go, most frequently invade the blood stream and it would appear often play the important role in many of the secondary conditions.

Chronic Infections

B. influenzæ is a frequent finding in the sputum of patients with chronic bronchitis, pulmonary tuberculosis and other chronic conditions in the respiratory tract. Boggs recovered this bacillus from two cases of bronchiectasis, Richards and Gurd had a similar case and Tedesko reported several. The literature is filled with references to the finding of B. influenzæ in cases of chronic bronchitis. Those reported by Lord, Madison and Tedesko quoted above will serve as examples. The frequent positive cultures in cases of pulmonary tuberculosis so often referred to in reviews of the literature and the significance of these findings, as pointed out by Scheller, are important as bearing on the much debated subject of the effect of influenza on this disease. These types of chronic infection by the influenza bacillus should be more generally recognized as they undoubtedly will become more numerous following this last epidemic if we can judge from the experience of the past.

Infections of the Pleura

The recovery of B. influenzæ from the pleural cavity is not uncommon as is shown in the above review. The findings of MacCallum, Cole and others during the spring of 1918 are particularly interesting. Beall in 1906 reported a case of empyema with large quantities of green pus in which B. influenzæ was found in pure culture.

Sinuses of the Head

Infection of the accessory sinuses of the head has long been recognized as occurring in influenza. Frankel found B. influenzæ in 4 from 40 infected antra. Lindenthal, who was particularly interested in the question of sporadic influenza, found the bacillus in one or more of the head sinuses in six of eight carefully studied cases. He considered that the B. influenzæ remained in these areas during inter-epidemic times and from hence caused the sporadic outbreaks of influenza. Howard and Ingersoll reviewed the literature up to 1898 and grew B. influenzæ from one of three acute antral diseases. They did not find it, however, in 12 chronic cases. Clemens believed the influenza bacillus to be present in the sinuses rather frequently in cases where it was overgrown or difficult to culture from the lower respiratory secretions. Moszkowski grew it in one case from the pus of the antrum. Tedesko recorded several positive results and many others are reported in the literature.

The two cases reported by Lacy (1918), the findings during the present epidemic by Stone and Swift of B. influenzæ in 13 of 28 sphenoidal and 6 of 12 ethmoidal sinuses cultured at necropsy, those by Spooner, Scott and Heath, of B. influenzæ in four frontal sinuses and in eight sphenoidal, and the recovery by Wolbach of B. influenzæ in cultures from the sinuses in certain cases where the lung cultures were negative, emphasize the importance and frequency of the infection by this organism in these cavities. Keegan, who laid particular stress on lung punctures and autopsy examinations, pointed out that in throat cultures the probability that the influenza focus is often not in the pharynx but in some recess of the nasal cavity.

H. E. Robertson in the spring of 1918 reported the infection of the sinuses in seven cases of tracheo-bronchitis with patches of broncho-pneumonia and the growth of B. influenzæ from sphenoid, ethmoid or frontal sinuses of all these cases. He also found this organism in the sphenoid of six cases dying with various diseases as well as in two accident cases with death under 24 hours. The importance of these results was laid stress on by the author, not only on account of the probable toxic absorption and the general menace of spread, but, more particularly, because such individuals, acting as carriers, could furnish foci for the spread of epidemics.

Eye and Ear

Infections of the eye by the influenza bacillus are quite common. This subject is fully discussed by Axenfeld (text-book, “The Bacteriology of the Eye”). Giani and Picchi found it in the eye in 66 per cent. of influenza cases, in 90 per cent. of epidemic conjunctivitis, and in the normal eye of 5.8 per cent. Wynekoop, in 1903, reported having found this organism in cases of conjunctivitis in 1899. Guiral, in the recent epidemic, found influenza bacillus constantly present in the secretions in cases of what seemed to be Week’s conjunctivitis. Ulceration of the cornea was rather common. One such case is mentioned in which there was no pain in the eyes, but general symptoms of influenza. The middle ear is also sometimes infected. Between the report of Kossel in 1893 and that of Stone and Swift in 1918, who found the middle ear and mastoid to contain B. influenzæ in 8 of 17 cases, there have been many references in the literature to this complication by the influenza bacillus. The evidence indicates, however, that in the middle ear, as in the pleural cavity, the secondary bacteria are far more often the important ones.

Meninges

Influenzal meningitis seems to stand by itself as a manifestation of the pathogenic effects of B. influenzæ. The literature is too voluminous to review in this place, but the evidence would seem to point to a more invasive and pathogenic type of this organism, if not to a separate member of the group.

Invasion of the Blood Stream

The evidence in clinical influenza would suggest at times a bacteremia in addition to the severe toxemia, which is such a constant feature of the disease. Simultaneously with the discovery of B. influenzæ, Canon reported finding bacilli of similar morphology in blood smears, but was unable to grow them, and it would appear at least doubtful that he was dealing with the influenza bacillus. Meunier is probably the first who grew this organism from the blood. He recovered it from 8 blood cultures out of 10 in cases of broncho-pneumonia following measles, and in one other case of broncho-pneumonia. A very full discussion of this question is to be found in Canon’s book on “The Bacteriology of the Blood in Infectious Diseases.” Of particular interest are the results of Ghedini, who made a careful study of 28 influenza patients. B. influenzæ was grown from the blood in 18 of these at the height of the fever, while in the 10 negative cases the disease was milder or the blood was taken only after the temperature had fallen. The amount of blood used was 20-30 c.c., and it was cultured in lecithin broth. In practically all of his cases several cultures were taken, and in a number of the positive cases negative results were obtained both before and after the acme of the fever. He also grew the bacillus from 8 of 14 spleen punctures of these patients. Madison (1910) reported the recovery of this bacillus from the blood of a patient with a primary broncho-pneumonia who recovered. This author also used about 30 c.c. of blood. Thursfield, in 1910, also reported two cases of B. influenzæ bacteremia in which the organisms were recovered at the height of the temperature. One had influenza, the other phlebitis, and both recovered. Tedesko and several others have found it in the heart’s blood in many cases, more especially in broncho-pneumonia after measles.

During the present epidemic the positive cultures of this bacillus from the blood have been rather infrequent. J. S. Fleming had 2; 2 are quoted in the report of the Influenza Committee of the Advisory Board to the D. G. M. S. (Peters and Cookson); Medalia had 2 during life and 19 of 34 at autopsy; Orticoni, Barbie and Leclerc in 5 of 10 blood cultures in one series, and 7 of 19 in another; Stone and Swift 2 at autopsy; McKeekin, in Australia, influenza-like bacilli in 4; Blanton and Irons three times in the heart’s blood, one of these pure; Spooner, Scott and Heath twice in the heart’s blood at autopsy, and Wollstein and Goldbloom from the heart’s blood in one child. In the majority of these findings the bacillus was not found in pure culture. Abrahams and his associates found the B. influenzæ along with a pneumococcus and M. catarrhalis from the heart’s blood in one case. In our positive blood culture there was evidence of the same mixture being present.

Before drawing sweeping conclusions against the invasion of the blood by B. influenzæ it must be remembered that the quantity of blood used has been generally only about 10 c.c., and often much less, the difficulty of observing growth if the culture is pure has been largely overlooked, the use of more favorable media than blood agar and the possible inhibitory action of influenzal blood, as suggested by Wittingham and Sims, Rivers and others, has not been considered, and further that sufficient care has not been exercised to obtain blood at the most favorable period in the disease. It may be recalled that the problem is quite similar to that of demonstrating the organisms in the blood in patients with streptococcus viridans bacteremia.

All the available evidence, however, points to the invasion of the blood in influenzal infections as being a very fleeting one. Unless this is true, it would be surprising in the many hundreds of blood cultures which have been taken in the concentrated study of patients during the recent pandemic, if more successful cultures had not been obtained. General infections with localization of B. influenzæ in different parts of the body are here of interest—such as that reported by Slawyk and others. Whether the strains causing meningitis, and which apparently more frequently invade the blood, are really different members of the hemophilic group or only forms with a higher invasive power is still, I believe, an open question.

Endocarditis

In endocarditis the B. influenzæ is probably, after streptococci, the organism most frequently isolated from the blood. Rosenthal from heart’s blood at autopsy, Schlangenhaufer, Jehle two cases, Horder (1907) six cases, and who believed he was the first to isolate B. influenzæ from the blood, Tedesko in a number at autopsy, Spat, F. J. Smith, Saathoff, Libman four cases, Sacquepee, McPhedran, Mann, Rainaford and Warren three cultures from two patients, and a number of others all bear witness to its frequency.

Other organs of the body are sometimes found to contain B. influenzæ. Adrian, Schultes, Basile and Tedesko have all recovered this organism from the diseased appendix. Several years ago a bacillus, considered, to be B. influenzæ, was grown from the pus of an appendix abscess in our laboratories. Wright found it in pyelonephrosis. Klieneberger found influenza-like bacilli in cases of cystitis. Menko reported the bacillus from orchitis, and Cohn found numerous influenza-like bacilli in the discharge from urethritis. Meunier found it in pure culture in a case of osteoperiostitis. Huyghe, Besancon and Griffon recovered it from infected joints, as did Pacchioni in a general infection. Weil found it in the pus about the hip joint one month after an attack of influenza. This short review serves to illustrate that the influenza bacillus, although generally limited to infections in the respiratory tract, is, nevertheless, capable of infecting other parts.

Immunity—Phagocytosis

Phagocytosis of the B. influenzæ has been very frequently noted in the study of sputum smears. It has been observed, moreover, that this phenomenon occurs most frequently when the patient is on the road to recovery (Pfeiffer, Martin, and others), and it may indicate an important reaction on the part of the body to this organism. Tunnicliff in a recent report, however, did not find the opsonic index to be raised above the normal in her patients, and Tunnicliff and Davis had difficulty with a spontaneous phagocytosis of this bacillus. This difficulty was to a large extent absent in her later study.

Agglutination

Agglutination tests have been used by many investigators in attempts to determine a specific reaction in the sera of persons suffering from influenza. Such reactions develop, as we know, against secondary infecting bacteria, so that unqualified conclusions cannot be drawn that agglutinins in the sera of patients against B. influenzæ indicate the etiological importance of this organism. Vagedes using a dilution of 1-50 found 8 positives among 27 patients tested. Lord found the test most inconstant. Ghedini obtained useful results by using serum in dilutions 1-20 to 1-30, and had 17 positives from 28 influenza cases. He found agglutinins present three to four days after the height of the infection, and noted that the sera became practically normal after three to four weeks. Fichtner, although he obtained agglutination with sera of influenza patients in high dilutions (1-100 and 1-750), found his controls were often agglutinated, and consequently drew no conclusions. Wollstein (1906) did a series of agglutination tests, using various strains of B. influenzæ. The sera of patients she found very unsatisfactory, but by immunizing rabbits with this organism she obtained sera with titres up to 1 in 400. She could find no differences among the various strains studied. Somewhat similar results were obtained by her in 1915 working with strains from the meninges and the respiratory tract. Odaira carried out a rather extensive series of tests, using immunized rabbit sera and a special method of making his bacterial emulsions. He was able to distinguish B. influenzæ from both B. pertussis and the so-called Cohen’s bacillus of meningitis. Friedberger’s dog bacillus, however, could not be differentiated from B. influenzæ by this means. A. Fleming during the recent epidemic had good results with the sera of 21 patients. He incubated at 50° C. for two hours. He also used sera of immunized rabbits and got marked agglutination against the homologous strain, but varying results with other strains. He noted some strains agglutinated readily, while others did not. Eyre and Lowe noted an increase in agglutinins in the sera of people vaccinated against the influenza bacillus. Couret and Herbert could distinguish two types and a possible third among their strains. Park and his co-workers found numerous types by means of agglutination. Absorption of agglutinins was found helpful by these last two workers. There are so many factors capable of altering the sensitiveness of bacteria to agglutination, as in the well-known experiments of Neufeld, that we must recognize that much work is still to be done before we can properly interpret the results of these agglutination tests.

Binding of Complement

Complement fixation tests were carried out by Odaira but his results were much less satisfactory than those he obtained by means of agglutination. Rapaport made an extensive study of this test, using the sera of patients in various stages of convalescence. Three hundred and fifteen convalescents showed 54.5 per cent. positive while 300 controls only gave 9.5 per cent. positive results. Most of the positive cases were in patients three to five days after their illness, but the reaction was found in convalescents after from 1 to 45 days. Sera from acutely ill patients at times showed negative or slightly positive reactions but these same sera after keeping for some days and retesting often gave strongly positive results. This would appear to be a promising field for investigation.

Anaphylaxis

Hypersensitiveness was noted by W. F. Robertson in chronic infections with B. influenzæ. Wollacott in a letter to the British Medical Journal suggested that the severity of the recent outbreak of influenza may possibly be due to the development of a state of anaphylaxis. There would seem to be at least some evidence in favor of such a view in the fact that the severe outbreak was preceded by epidemics of a milder form of influenza and that the influenza bacillus was probably widely spread during this time. Greenwood, as quoted above, noted that primary cases always precede the mass attack. Of course, the term anaphylaxis has been used to explain almost everything. Nevertheless, the theory is interesting. The skin tests which we did for hypersensitiveness were, as I have noted above, negative but there is a possibility that the failure of the reaction may indicate a higher resistance or even an antitoxin, now that the bacillus can be classed as a toxicogenic one. Anti-influenza sera have been produced by a few investigators (Latapie, Wollstein) but have not found any practical application during this pandemic. Vaccination is discussed elsewhere in these studies.

Experiments on the Human

There has never been in the history of medicine so many experiments on human beings as have been carried out in the attempts to discover the etiological factor in the recent pandemic of influenza. Davis has called attention to a successful human inoculation with pure cultures of B. influenzæ which he performed in 1906. During the present investigation at least 200 men have volunteered as experimental subjects, and the results of many different methods of attempting to transmit the disease, have been disappointing and inconclusive. I will not attempt to review the reports at present available, as a great deal of the work done has not yet appeared in print. The important point is that the results do not affect the various views held as to the causative agent in pandemic influenza nor the massive evidence for transmission of the disease under natural epidemic conditions.

It is my opinion, as expressed above, that practically all of the population are rapidly infected during such a pandemic as we have had. The resistant have escaped, and it would appear to be very difficult to break down this resistance. The human experiment carried out by Pettenkofer on himself and his assistant with vibrion choleræ is an example, but we have numerous others demonstrating the same kind of phenomena in most of our diseases of established bacterial origin. In diphtheria we have an explanation in the varying antitoxic content of the sera, but we really know very little of what are the actual factors in preventing or determining infection among exposed individuals in the natural history of most diseases. The reports of Leonard Hill and Gregor are well worth reading in this connection, as well as the editorial in the same number of the British Medical Journal. We are not in a position to be very dogmatic on the causes of epidemics. The mere presence of the bacteria or any other living virus is not in itself sufficient to explain the phenomenon, and one of the chief objects of this paper is to indicate from the collected facts, that in the words of Flexner, “the case against the influenza bacillus is not proved.”

Conclusions

1. B. influenzæ is one of a group of hemophilic bacteria and there are probably strains of this organism which may be differentiated which will lead to further subdivisions of the group.

2. B. influenzæ as we understand it today, is distinguished by its morphological and staining characters; its requiring hemoglobin in some form for its development; its showing symbiotic reactions with other bacteria which stimulate its growth; the production of a toxine and its usual low pathogenicity for animals.

3. The media found most favorable for its growth are those containing blood with the hemoglobin content altered in certain ways, (1) by heating, (2) the addition of various chemicals, (3) by the action of other bacteria or their products. The heated blood agar I have found to be a most efficient and readily prepared medium.

4. Since B. influenzæ is so difficult to isolate, it is necessary to be very cautious in interpreting results unless the greatest effort has been made to demonstrate the presence of this organism.

5. B. influenzæ should be considered, from the evidence at hand, as the bacterial causative agent in epidemic influenza, and it should be recognized that secondary infections following the primary attack by this organism are both frequent and important. This view I believe the logical one, unless much more convincing evidence than we have today may demonstrate another more probable living virus as the cause.

6. B. influenzæ is a frequent etiological factor in purulent and chronic bronchitis, broncho-pneumonia and other acute and chronic respiratory infections, in meningitis, endocarditis, sinusitis, conjunctivitis and other conditions, as well as in complications of many other diseases.

7. There are many carriers of the bacillus among our population, both in apparently normal individuals and in those suffering from chronic infections of bronchi, sinuses or other parts.

8. The problem of what constitutes resistance or susceptibility to this infection are as far from solution as they are in most other respiratory diseases, and the attempts to explain the reasons for epidemics have been as futile as they are for meningitis and many other respiratory epidemics.

9. It would not appear that the immunological reaction against this infection has been discovered, but the possibility of its being of an antitoxic nature opens an interesting field for investigation.