Epidemic Respiratory Disease

CHAPTER III SECONDARY INFECTION IN THE WARD TREATMENT OF INFLUENZA AND PNEUMONIA — Secondary Infection with S. Hemolyticus in Pneumonia — Secondary Infection with Pneumococcus in Pneumonia — Secondary Contact Infection in Influenza — Methods for the Prevention of Secondary Contact Infection in Influenza and Pneumonia — Summary

One of the most pressing problems that presented itself in the care of influenza and pneumonia patients in the army cantonments during the recent epidemic was the danger of secondary contact infection because of the overcrowding of the base hospitals, nearly all of which were taxed far beyond the limits of their capacity. That this danger was very real was fully demonstrated by certain studies in ward infection that this commission was able to make at Camp Pike^[49]. It is the purpose of the present section of the report to present these studies and to discuss the means whereby this danger may be most successfully met.

It is perhaps well, first to define exactly what is meant by secondary contact infection in influenza and pneumonia. In our experience at Camp Pike it was found that a very large percentage of the pneumonias following influenza were accompanied by secondary infection with pneumococcus, some few being caused by hemolytic streptococcus. The types of pneumococcus encountered were almost entirely those that are found normally in the mouths of healthy men, approximately 85 per cent being Types II atypical, III, and IV. It has been generally accepted that infection with these types of pneumococci is usually autogenous—that is, that under the proper conditions of lowered resistance an individual becomes infected with the pneumococcus that he carries in his own mouth. Many observations 84made during the course of the present work have suggested that this is probably not so in many instances and that the influenza patient may not be so much in danger from the pneumococcus that he normally carries in his own mouth as he is from that carried by his neighbor, in other words, he is in danger from contact infection. The same considerations hold true with respect to hemolytic streptococcus infection. Secondary contact infection in cases of already existing pneumonia following influenza were found to occur frequently. These were for the most part caused by hemolytic streptococcus infection superimposed upon a pneumococcus pneumonia. Many instances of double pneumococcus infection, however, either coincident with or following one another were encountered.

Pneumonia caused by streptococci was repeatedly observed^[50] during the pandemic of influenza which occurred in 1889–90. With clearer recognition of the characters which distinguish varieties of streptococci several observers have shown that secondary infection with hemolytic streptococci may occur during the course of pneumonia and though definite evidence has been lacking have suggested that it may be acquired within hospital wards. That a similar secondary infection with S. hemolyticus in pneumococcus pneumonias following influenza occurred not infrequently at Camp Pike during the epidemic was shown by bacteriologic studies made during life and at autopsy in a considerable series of cases. During the early days of the epidemic of influenza, secondary streptococcus infection was almost entirely limited to certain wards which were opened for the care of the rapidly increasing number of patients with pneumonia. During this period these wards were overcrowded, organization was incomplete, and the opportunities for transfer of infection from patient to patient 85were almost unlimited. The spread of streptococcus contagion and its fatal effect may be clearly brought out by comparison of these wards with wards that had long been organized for the care of patients with pneumonia.

Ward 3 had been in use for the care of patients with pneumonia for some time prior to the outbreak of influenza. It was provided with sheet cubicles and conducted by medical officers, nurses and enlisted men accustomed to the care of patients with pneumonia, ordinary precautions being taken against transfer of infection from one patient to another. The data in Table XVII show the average number of patients in the ward, the number of new cases of pneumonia admitted, and the number of deaths among patients admitted during the corresponding period, for three periods of ten days each from September 6 to October 5. The types of infection in fatal cases as determined by cultures taken at autopsy are also shown.

During the period from September 6 to 15, just prior to the outbreak of influenza in epidemic proportions, the ward had an average population of 18.6 patients. The total number of new patients admitted was 11, of whom 3 died, a mortality of 27.2 per cent. All these cases were pneumococcus pneumonias as determined by bacteriologic examination of the sputum at time of admission. The 3 fatal cases showed pneumococcus infection at autopsy. During the second period, from September 16 to 25, with the outbreak 86of the epidemic of influenza, the ward rapidly filled with new cases of pneumonia, attaining an average population of 46.1 patients. Of the 52 new cases admitted 16 died, a mortality of 30.7 per cent. Again all the cases admitted during this period in which bacteriologic examination of the sputum was made, were found to be pneumococcus pneumonias with one exception. This case, admitted on September 21 and dying two days later, had a hemolytic streptococcus pneumonia. Fortunately, though quite by accident, he was placed in a bed at one end of the porch and no transmission of streptococcus infection to other cases in the ward took place. At autopsy 13 cases showed pneumococcus infection; the foregoing case, hemolytic streptococcus. During the third period from September 26 to October 5 the ward became even more crowded, having an average of 58.6 patients; 23 new cases were admitted, 8 of whom died, a mortality of 34.7 per cent. Autopsy showed that 5 of these were pneumococcus pneumonias and 1 was caused by hemolytic streptococcus infection. It is noteworthy that the death rate from pneumonia gradually increased as the ward became more and more crowded. This may possibly be attributed in part to the increasing severity of the pneumonia during the early days of the influenza epidemic. That it was in part directly due to secondary contact infection with pneumococcus will be shown when the transmission of pneumococcus infection is discussed. In spite of the overcrowding of the ward the introduction of 2 cases of streptococcus pneumonia did not cause an outbreak of streptococcus infection. Whether this was due to precautions taken against the transfer of infection or was merely a matter of good luck is difficult to say, in view of the fact that a considerable amount of transfer of pneumococcus infection from one patient to another did occur.

Ward 8 had long been used for the care of colored patients with pneumonia. As in Ward 3 cubicles were in use 87and ordinary precautions against the transfer of infection were used. The data are presented in Table XVIII.

During the period from September 6 to 20, prior to the outbreak of influenza in epidemic proportions among the colored troops, the ward had an average population of 25.5 patients; 18 new cases of pneumonia were admitted during this period, all of whom were pneumococcus pneumonias as determined by bacteriologic examination of the sputum at time of admission to the ward. Only 2 died, a mortality of 11.1 per cent, autopsy cultures showing pneumococcus in both cases. All these patients were treated on the porch of the ward while they were acutely sick. During the second period from September 21 to October 5, when the influenza epidemic was at its height, the ward rapidly filled with active cases of pneumonia and became distinctly crowded. It contained an average of 46.1 patients, but had actually reached a population of 64 patients at the end of the period. Of the 59 new cases admitted, 20 died, a mortality of 33.9 per cent, 10 with pneumococcus pneumonia, one with hemolytic streptococcus pneumonia. In 9 there was no autopsy. The conditions in Ward 8 were quite analogous to those in Ward 3. In spite of the overcrowding during the second period no outbreak of secondary infection with S. hemolyticus occurred, but secondary pneumococcus infection did occur as will be shown below.

88In contrast with these two wards are Wards 1 and 2 in which widespread secondary contact infection with S. hemolyticus took place. Ward 2 was opened September 26, at the beginning of the period when 20 new wards for pneumonia were organized. From September 26 to October 1 the cubicle system was not in use, the ward was crowded, organization was imperfect, and few precautions were taken to prevent transfer of infection from one patient to another. On October 2 the cubicle system was installed and precautions against transfer of infection were instituted. The data are shown in Table XIX.

During the first three days 40 patients with pneumonia were admitted to the ward. Of these 40 patients, 27 died, a mortality of 67.5 per cent. Cultures at autopsy showed that 23 of these died with hemolytic streptococcus infection, none of pneumococcus infection. In four there was no autopsy. To appreciate the full significance of these figures it must be emphasized that these patients at time of admission to the ward in no way differed from those admitted to Ward 3 during the corresponding period and were not in any sense selected cases. The type of infection in 9 of these patients had been determined by bacteriologic examination 89of the sputum just prior to or immediately after admission to the ward before opportunity for secondary contact infection in this ward had occurred. All 9 were shown to have pneumococcus pneumonia free from hemolytic streptococci at that time. All 9 died, 7 with secondary streptococcus infection as shown by cultures taken at autopsy, 1 with a secondarily acquired Pneumococcus Type III infection—sputum showed a Pneumococcus Type IV on admission—and in 1 there was no autopsy. In view of the fact that bacteriologic examination of the sputum in cases of pneumonia following influenza had shown that the large majority of them were due to pneumococcus infection, it is probable that most of the other cases of pneumonia admitted to this ward were pneumococcus pneumonias at time of admission, and that they acquired the streptococcus infection after admission.

During the next three days 17 new patients were admitted, of whom 6 died, a mortality of 35.3 per cent. Cultures at autopsy showed pneumococcus infection in 2, streptococcus in 2. It is noteworthy that the porch was first put into use on September 29. Of the 12 patients admitted on this date, 8 were treated throughout the acute stage of their illness on the porch. Of these 8 patients but one died, of a Pneumococcus Type IV infection and none became infected with S. hemolyticus. From October 4 to October 6, 10 patients were admitted, of whom 4 died. Cultures at autopsy showed pneumococcus infection in 2, hemolytic streptococcus in 1.

The widespread prevalence of hemolytic streptococcus infection in this ward as compared with its almost entire absence in Wards 3 and 8 is very striking. Cultures made during life and at autopsy have shewn clearly that it was due to rapid spread of contagion throughout the ward. The almost unlimited opportunities for transfer of infection from patient to patient, during the first six days the ward was in use, undoubtedly greatly facilitated this spread. 90From the data available it is impossible to state exactly when and by which patients hemolytic streptococcus infection was introduced into the ward, but it must have been very early since the death rate was very high from the beginning, and the first 23 cases coming to autopsy died with streptococcus infection.

Ward 1 was opened on September 24. From that date until October 2 no cubicles were in use and few precautions were taken against transfer of infection. On October 2 cubicles were installed and ordinary precautions to prevent transfer of infection were instituted. On October 6 the ward was closed to further admissions. The data presented in Table XX are divided into two periods, because on September 29 and 30, 4 patients with streptococcus pneumonia were admitted to the ward.

During the first period from September 24 to 29 the ward contained an average of 35.8 patients, being only moderately crowded; 34 cases of pneumonia were admitted, of whom 11 died, a mortality of 32.3 per cent. It is noteworthy that deaths among this group which occurred prior to September 30 were due to pneumococcus infection with one exception, a patient entering the ward on September 26 and dying the following day. Of the other 2 patients in this group who died with hemolytic streptococcus pneumonia, 1 was admitted to the ward on September 25, was 91shown to be free from S. hemolyticus on September 30, and died on October 12 with a secondarily acquired streptococcus pneumonia and empyema; the other was admitted on September 29 with streptococcus pneumonia and died the following day.

During the second period from September 30 to October 5 the ward contained an average of 55.3 patients, being very overcrowded; 40 new cases of pneumonia were admitted of whom 24 died, a mortality of 60 per cent. Cultures taken at autopsy showed that 6 died of pneumococcus pneumonia, 14 with hemolytic streptococcus infection. As in Ward 2, patients admitted to this ward were in no way selected and were probably, as experience has shown, in large part pneumococcus pneumonias at time of admission. The widespread dissemination of hemolytic streptococcus and its fatal effect following the introduction of the organism on September 29 and 30 is only too evident.

The foregoing studies have shown that hemolytic streptococcus infection may spread by contagion throughout an entire ward with great rapidity. Other observations have demonstrated that pneumococcus infection may be transmitted in the same way. Only three instances of this nature will be cited. The first occurred in Ward 3 (Table XXI). Between September 6 and 16 no cases of pneumonia caused by Pneumococcus Type II had been admitted to the ward. On September 16 Pvt. Swain was admitted to the ward and placed in Bed 3. Bacteriologic examination of his sputum showed that his pneumonia was caused by Pneumococcus Type II. At this time Pvt. Pullam, who had been admitted to the ward on September 9 with a pneumococcus Type IV pneumonia, occupied Bed 5 separated from Bed 3 by one intervening bed. He had had his crisis on September 14 and was doing well, his temperature being normal. On September 24 he developed a second attack of pneumonia and died on September 30. Cultures at autopsy showed Pneumococcus Type II in heart’s blood and lung, Pneumococcus Type II and B. influenzæ in the right bronchus. Pvt. Wolfe was admitted to the ward with bronchopneumonia on September 17 and placed in Bed 6 next to Pvt. Pullam. Pneumococcus Type IV and B. influenzæ were found in his sputum. His temperature had fallen to normal by lysis on September 21 and he was doing well. On September 23 his temperature suddenly rose and he developed a second attack of pneumonia. Pneumococcus Type II was isolated by blood culture on this date. He recovered. In both instances Pneumococcus Type II was acquired after the admission of a patient with a Pneumococcus Type II pneumonia, the opportunity for contact infection having been favored by the association of these patients in adjoining beds.

The second instance is almost identical and occurred on the opposite side of Ward 3 at about the same time (Table XXII). Pvt. Linehan was admitted on September 16 and placed in Bed 30. Pneumococcus Type IV was found in his sputum. Pvt. Thompson was admitted the following day with a Pneumococcus II atypical pneumonia and placed in Bed 28. The next day Pvt. Smith was admitted and placed in Bed 26. Pneumococcus Type II was found in his sputum. On September 19 Pvt. Thompson recovered by crisis and was doing well. On September 21 he had a chill, his temperature rose to 104.4° F. and he developed a second attack of pneumonia. He died on September 29; cultures at autopsy showing Pneumococcus Type II in heart’s blood and left pleural cavity, Pneumococcus Type II and B. influenzæ in bronchus and lung. Pvt. Linehan had begun to improve on September 24 and his temperature was falling by lysis. On September 26 he became worse, developed signs of pericarditis and died on September 30. Cultures from lungs and bronchus at autopsy showed Pneumococcus Type II and B. influenzæ. In both instances the fatal secondary infection with Pneumococcus Type II was undoubtedly acquired from Pvt. Smith in the nearby bed.

The third instance occurred in Ward 8 (Table XXIII). Pvts. Lewis and Scott were admitted on September 21 and were placed in adjoining beds, 50 and 51. Lewis showed Pneumococcus Type I in his sputum, Scott Pneumococcus II atypical. The following day Pvts. Pighee, Jones, and Columbus were admitted and given Beds 48, 49 and 53 respectively. All showed Pneumococcus II atypical in the sputum. Pvt. Lewis with Pneumococcus Type I pneumonia recovered by crisis on September 29. His temperature remained normal until October 2 when it suddenly rose to 104.2° F. He developed a second attack of pneumonia and died on October 8. Cultures at autopsy from heart’s blood and lung showed Pneumococcus II atypical, from the bronchus Pneumococcus II atypical and B. influenzæ. It is, of course, impossible to say from which one of his neighbors Pvt. Lewis acquired his second pneumococcus infection.

It is noteworthy that these instances of secondary contact infection with pneumococci occurred in wards where every precaution was supposedly taken to prevent transfer of infection from one patient to another. It is true however that the wards were greatly overcrowded at the time. Many other instances of secondary pneumococcus infection in cases of pneumonia following influenza were encountered in which it was impossible to trace the source of infection, many combinations of different types of pneumococcus being found. There were two instances in which Pneumococcus Type IV was found in the sputum by inoculation of white mice shortly after onset of pneumonia, whereas secondary infection with other types was found at autopsy, one with Pneumococcus Type II, one with Pneumococcus Type III. In 2 cases by inoculation of white mice, two types of pneumococcus were found simultaneously in the sputum coughed from the lung, in one Pneumococcus Types III and IV, in the other Pneumococcus Types I and IV. There were 5 cases in which two types of pneumococcus were found in cultures at autopsy as shown in Table XXIV. Combined pneumococcus infections of this nature are almost never encountered in pneumonia occurring under normal conditions in the absence of epidemic influenza.

The foregoing data show that infection with one type of pneumococcus may readily be superimposed upon or closely follow infection with another type. Cases have been cited in which it was clearly demonstrated that this was due to contact infection. It is furthermore evident that pneumonia caused by one type of pneumococcus affords no reliable immunity against pneumonia caused by another type. The same conditions that favored the spread of hemolytic streptococcus infection also favored the transfer of pneumococcus infection from patient to patient.

The material so far presented has dealt with contact infection in cases of pneumonia following influenza. That a similar contact infection in cases of influenza treated in crowded hospital wards is responsible in considerable degree for the development of pneumonia in cases of influenza seems quite probable. It has already been stated that this pneumonia was found in large part to be caused by 96infection with types of pneumococcus that are found in the mouths of normal individuals. It has been fairly definitely established by Stillman^[52] that lobar pneumonia caused by pneumococcus Types I and II is in all probability due to contact infection, and definite instances of such infection by Pneumococcus Type II have been reported above. In a recent communication Stillman^[53] has furthermore shown that of the various groups of Pneumococcus II atypical those most frequently associated with pneumonia are rarely found in normal mouths, while those infrequently associated with pneumonia are more commonly found. Whether similar considerations will hold true for pneumococci of Group IV can only be determined by further investigation. It has been stated that certain observations made during the course of this work have suggested that cases of pneumonia which complicate influenza may be due to contact rather than to autogenous infection. The data available are far too limited to establish this fact and it would require a very extensive study to furnish conclusive evidence.

Certain general observations have suggested this point of view. It is well recognized that the incidence of pneumonia in patients with influenza has been much higher where overcrowding has existed. It would seem probable that this has been in large part due to the greater opportunity for the dissemination of organisms capable of producing pneumonia and the consequently increased opportunity for secondary contact infection among patients treated under such conditions. The not infrequent occurrence of influenzal pneumonia due to combined infections of the different types of pneumococci, hemolytic streptococci, staphylococci, and other bacteria, instances of which have been cited, is in harmony with this view, especially since pneumonia under ordinary conditions is rarely found to be associated with mixed infections of this nature. It 97is true that healthy individuals occasionally carry more than one type of pneumococcus simultaneously in the mouth, though this is very infrequent, and autogenous infection occurring in such individuals might account in some instances for the mixed pneumococcus infections encountered. By way of analogy it has been clearly shown in other studies by the Commission on the relation of hemolytic streptococcus carriers to the complications of measles, that secondary infection of the respiratory tract with S. hemolyticus is in very large part due to contact infection, the chronic carrier rarely developing complications due to this organism.

To obtain further light on this question the type of pneumococcus present in the mouths of 46 consecutive cases of early uncomplicated influenza was determined by the mouse inoculation method at time of admission to the receiving ward of the hospital before the patients had been associated, with the purpose of determining if cases among this group which subsequently developed pneumonia might be shown to have acquired a pneumococcus which they did not carry at time of admission. This group of patients was treated in a special ward set apart for the purpose. The patients were assigned to beds in rotation and confined in bed until thoroughly convalescent. Beds were well separated and cubicles, masks and gowns were in use. Cultures were made from the ward personnel. By these procedures an accurate record was kept of all sources of pneumococcus infection. The types of pneumococcus found in the mouths of these patients at time of admission are shown in Table XXV.

Only 1 patient in this group developed pneumonia. At time of admission he had no pneumococcus in his mouth as determined by inoculation of a white mouse with his sputum. Examination of the sputum by the same method at time of onset of pneumonia three days after admission showed Pneumococcus Type III. The only ascertainable source of infection in this case was one of the ward attendants who carried Pneumococcus Type III in his throat in sufficiently large numbers to be demonstrable by direct culture and who frequently came in contact with the patient. In this instance the development of pneumonia was probably due to contact infection. An extensive study of this nature would be necessary to determine in what proportion of cases pneumonia following influenza is caused by secondary contact infection and in what proportion to autogenous infection. It is at least evident that contact infection with a type of pneumococcus found in the mouth of normal individuals may occur in influenza and be responsible for the development of pneumonia. Therefore every precaution should be taken to prevent it.

The methods at present in vogue for preventing the spread of contagion in wards devoted to the care of patients with influenza and pneumonia may be briefly enumerated: The separation of patients by means of sheet or screen cubicles, the wearing of masks and gowns by the ward personnel and to some extent by convalescent patients who are up and about the ward, and in some hospitals the separation of streptococcus carriers from noncarriers as determined by throat culture at time of admission. That these methods are of some value in preventing spread of infection cannot be denied, and it is probable that they are fairly 99effective under ordinary conditions when conscientiously carried out. That they inevitably break down in the presence of an overwhelming epidemic when hospital wards become overcrowded is only too evident. Under such conditions the sheets hung between the beds are constantly being displaced and are slight proof against a patient’s curiosity as to the identity and condition of the man in the adjoining bed; masks cannot be worn by patients seriously ill with pneumonia, during the very time when they are most dangerous and in greatest danger and those worn by the ward personnel are very rarely sufficiently well made to prevent spread of contagion by droplet infection as the studies of Haller and Colwell^[54] and Doust and Lyon^[55] have shown; the separation of streptococcus carriers from noncarriers as at present carried out cannot keep pace with the ever increasing influx of patients nor with the rapidity of the spread of the hemolytic streptococcus, in part because of the time required to make the bacteriologic diagnosis, in part because the amount of work involved cannot be accomplished by the laboratory personnel available. That this is so will be shown in data presented below. Not only do these methods break down in the face of an epidemic, but they often provide a false sense of security.

In searching for a solution of the problem it is essential to have the following considerations clearly in mind. Every patient with influenza must be considered a potential source of pneumococcus or hemolytic streptococcus infection for his neighbor until he is proved otherwise by bacteriologic examination. Every person engaged in the care of patients with respiratory diseases must also be regarded as a potential source of danger. Pneumonia cannot be regarded as one disease but must be looked upon as a group of different diseases, with more or less similar physical signs and symptoms, it is true, but caused by a considerable variety of bacteria, infection with any one of which not 100only provides no protection against infection with another, but may even render the individual more susceptible to secondary infection. Therefore, every patient with pneumonia must be regarded as an actual source of danger to his neighbor, at least until it is established that he has the same type of infection. All these considerations are especially true in the presence of influenza, for it has become evident that many organisms readily gain access to the lung and produce pneumonia in patients with influenza which under ordinary circumstances fail to cause disease of the respiratory organs.

Since secondary infection in respiratory disease is undoubtedly spread in large part by droplet and contact infection, the prevention of secondary infection must depend upon the elimination of these methods of transmission. Three solutions present themselves: (1) Ward treatment with absolute elimination of overcrowding and much wider separation of patients than has hitherto been deemed necessary; (2) segregation of patients according to type of bacterial infection; (3) effective individual isolation of every patient.

It has been clearly shown that treatment of influenza and pneumonia in overcrowded wards even with the use of such precautions to prevent transfer of infection as cubicles, masking of attendants and patients, etc., is attended by serious danger of contact infection and that such infection will almost inevitably occur. This is not at all surprising when it is remembered that we are treating in the same ward, in the case of pneumonia, a group of what are in reality entirely different diseases, all of which may be transmitted from one patient to another, and in the case of influenza a group of individuals who carry a variety of potentially pathogenic bacteria. No one would expect to treat cases of scarlet fever, measles, and diphtheria together in a hospital ward without having contact infection result. Among patients ill with influenza and postinfluenzal 101pneumonia, certainly streptococcus pneumonia and to some extent pneumococcus pneumonia may be transmitted quite as readily as any of these diseases. In view of these considerations it must be apparent if ward treatment of these diseases is to be continued without respect to type of bacterial infection, not only that overcrowding is absolutely contraindicated but also that much wider separation of patients than has hitherto been regarded as necessary is imperative. Furthermore, beds should be separated by permanent cubicles that cannot readily be displaced. Patients should be confined to their cubicles until thoroughly convalescent and when up and about should not be allowed to enter cubicles occupied by patients still sick. Medical officers, nurses and attendants who come into contact with the patients should use the same rigid precautions that are used in the care of patients with typhoid or erysipelas or meningitis with the additional use of means to prevent droplet infection of the patients, always bearing in mind that the respiratory tract in patients with influenza or postinfluenzal pneumonia is as susceptible to secondary infection as the postpartum uterus or an open surgical wound. In other words, the most rigid aseptic technic should be maintained. The recognition of a case of streptococcus pneumonia in a ward should be an indication for immediate quarantine of the ward until it has been shown by bacteriologic examination that there is no longer danger of spread of streptococcus contagion. This is done in the case of meningitis or diphtheria, neither of which diseases is comparable with streptococcus pneumonia in rapidity of spread or in resulting fatality.

Segregation of patients in wards according to type of bacterial infection while theoretically an improvement over the indiscriminate mixing of patients with many different types of infection presents many practical difficulties which make it impossible to carry out in the presence of an overwhelming epidemic. It is quite obvious that grouping of 102influenzal patients on the basis of the types of pneumococci that they carry in their mouths is impossible since the great majority of mouth pneumococci belong to Group IV and comprise a heterologous immunologic group. The separation of influenza patients who carry S. hemolyticus from those who do not would appear to offer a more hopeful field. Since we cannot make an immediate distinction between streptococcus carriers and noncarriers by inspection of the patient, this procedure requires the taking of throat cultures at time of admission to the hospital, the holding of patients for eighteen to twenty-four hours in receiving wards until the bacteriologic diagnosis has been made, and their subsequent distribution to streptococcus and nonstreptococcus wards. This is feasible when the admission rate is low and the number of streptococcus carriers found at time of admission is small. In the presence of an influenza epidemic it immediately becomes impossible to carry out in base hospitals as now constituted, since the demand for beds under such conditions at once converts a large part of the hospital into a group of receiving wards with little room remaining for subsequent separation of patients. The amount of bacteriologic work involved at once becomes prohibitive and the time required to make the bacteriologic diagnosis defeats its purpose since it allows the spread of hemolytic streptococcus to occur in the receiving wards during the interval.

The foregoing statements are based on results obtained in an attempt to separate streptococcus carriers from noncarriers in a limited group of cases of influenza at Camp Pike, the investigation being conducted during a secondary wave of influenza between November 27 and December 5. A special group of five wards consisting of one receiving ward and four distributing wards were set aside for the study. Cubicles, masks and gowns were in use and the wards were not crowded. The personnel on these wards did not carry S. hemolyticus in their throats. Patients entering 103the receiving ward were assigned to beds in rotation. Throat cultures were made on blood agar plates at time of admission. The plates were examined promptly the next morning, the diagnosis of S. hemolyticus being made by the characteristic hemolytic colonies and microscopic examination of stained smears. By this method a report reached the receiving ward at 9:30 a.m. and patients were promptly evacuated to the streptococcus and nonstreptococcus wards, where they were again assigned to beds in rotation, remaining confined in bed until convalescent. Confirmation of all strains of hemolytic streptococcus was subsequently carried out by isolation in pure culture, bile solubility test, and hemolytic test with washed sheep corpuscles. All cases free from hemolytic streptococci at time of admission who were sent to the “clean” wards were recultured daily throughout the period of study, those acquiring a hemolytic streptococcus being transferred to a streptococcus ward as soon as the bacteriologic diagnosis was made. The results are shown in Table XXVI.

One hundred and sixty-five cases were admitted to the receiving ward during the period of study as cases of influenza. 104Of these, 137 had influenza; 4 of those with influenza had pneumonia at time of admission, 23 had acute follicular tonsillitis, 3 epidemic cerebrospinal meningitis, 1 scarlet fever, and 1 Vincent’s angina. Sixty-six cases (40 per cent) showed hemolytic streptococcus in the throat at time of admission and were sent to the streptococcus wards; 99 cases (60 per cent) were negative for hemolytic streptococcus on admission, and of these 91 were sent to the “clean” influenza wards. Twenty-four of these clean cases subsequently became positive for S. hemolyticus. It is especially noteworthy that 14 of them acquired a hemolytic streptococcus during the short period that they were held in the receiving ward awaiting the report of the culture taken at time of admission, the first culture taken shortly after admission to the “clean” wards being positive. This result was undoubtedly due to the fact that these cases were unavoidably associated in the receiving ward with many carriers of hemolytic streptococcus. It is evident that cases which were supposedly free from streptococci but which in reality had picked up the organism in the receiving ward were constantly being sent to the “clean” wards. It is furthermore evident that if the precaution had not been taken of reculturing all clean cases on day of admission to the “clean” wards and daily thereafter these wards would soon have become saturated with hemolytic streptococci. Even under these conditions, 10 cases, after varying periods in the “clean” wards, acquired the organism in their throats. When it is stated that it required the full time of two men under very special conditions to carry out this work in a very limited number of cases and that it failed to keep “clean” wards free from hemolytic streptococci, it is only too apparent that the efficient separation of carriers from noncarriers in the presence of an epidemic of influenza is an impossible task.

The segregation of pneumococcus pneumonias following influenza according to type of infection is obviously impossible, 105since they are caused by an almost unlimited variety of immunologic types as far as present knowledge goes.

Even the efficient separation of streptococcus pneumonias from pneumococcus pneumonias would require a considerable team of workers and the closest cooperation between laboratory and ward staffs, so that no case of pneumonia would be sent to a pneumonia ward until the bacteriologic diagnosis had been made. In our experience this is rarely considered feasible even under ordinary conditions, and in the presence of an epidemic is nearly impossible because of the volume of work involved and the delay necessitated by bacteriologic methods. It is, nevertheless, absolutely essential if highly fatal ward epidemics of streptococcus pneumonia are to be prevented.

In view of the considerations discussed above, it is believed that the clear and most fundamental indication for the management of epidemic respiratory diseases in the army is to scatter patients as widely as possible instead of following the time-honored custom of concentrating them. In brief, abandon open ward treatment and adopt effective individual isolation of every case, maintaining as strict a quarantine as is demanded in other highly contagious and infectious diseases. The adoption of a strict aseptic technic in the handling of these patients is an evident corollary. Only by this means can the serious and highly fatal secondary hospital infections, which occur in influenza and pneumonia when these diseases are present in epidemic form, be prevented.

The prevention of secondary infection, prior to admission to the hospital, is another and more difficult problem. That opportunity for secondary contact infection in cases of influenza before patients reach the hospital is great seems unquestionable, since many cases have already developed these infections at time of admission. During the epidemic patients were crowded in regimental infirmaries, in ambulances, and in the receiving office of the hospital 106with every opportunity for droplet infection present. No study has been made of this question, but it seems reasonable that the same methods of prevention should apply, namely, effective separation of patients.

It is not within the scope of this paper to discuss details of method, but anything that is possible becomes feasible as soon as sufficient evidence can be brought to bear that it is a necessity. In the present instance it would seem that any means that can be used to reduce materially the terrific toll taken by respiratory diseases is an absolute necessity.

1. Secondary contact infection with pneumococci not infrequently occurs in patients with pneumonia following influenza when they are treated in hospital wards.

2. Secondary contact infection with S. hemolyticus readily occurs in patients with pneumonia and may spread rapidly throughout an entire ward with highly fatal results.

3. Secondary contact infection may be responsible for the development of pneumonia in patients with influenza.

4. Ward treatment of these diseases is fraught with serious danger which is greatly increased by overcrowding, by imperfect separation of patients by cubicles, and by imperfect aseptic technic of medical officers, nurses, and attendants.

5. It is probable that secondary contact infection can be effectively prevented only by individual isolation and strict quarantine of every patient.

CHAPTER IV THE PATHOLOGY AND BACTERIOLOGY OF PNEUMONIA FOLLOWING INFLUENZA — Bronchitis — Lobar Pneumonia — Bronchopneumonia — Peribronchial Hemorrhage and Pneumonia — Suppurative Pneumonia With Necrosis and Abscess Formation — Interstitial Suppurative Pneumonia — Suppurative Pneumonia with Multiple Clustered Abscesses Caused by Staphylococci — Empyema, Pericarditis and Peritonitis — Bronchiectasis — Unresolved Bronchopneumonia

Many observers have described isolated phases of the recent epidemic and of past epidemics of influenza. Few have had an opportunity to follow the pathology of influenza from the onset of an epidemic through a period of several months and to observe the succession of acute and chronic changes which occur in the lungs. Our commission arrived on September 5, 1918, at Camp Pike two weeks before the outbreak of influenza. The commission had previously made a careful study of the clinical course, the bacteriology and to a limited extent the pathology of pneumonia occurring at Camp Funston where there was little if any influenza. Study of the records preserved at the base hospital at Camp Funston had convinced us that this camp had passed through an epidemic of influenza during the spring of 1918, this epidemic being followed by a very severe outbreak of pneumonia. Our investigation at Camp Funston had brought to our attention those phases of pneumonia which with the facilities of a base hospital laboratory could be most profitably studied with a view to determining the causation, the epidemiology and the prevention of the pneumonias prevalent in the American army.

Study of pneumonia after death offers the only opportunity of determining the relation of pulmonary lesions to the considerable variety of microorganism associated with them. Clinical diagnosis furnishes no certain criterion for distinguishing lobar and bronchopneumonia; suppurative 108pneumonia is rarely recognizable during life. The relation of pneumococci, streptococci, staphylococci or B. influenzæ to one or other type of pneumonia can be determined with accuracy only after death; for the demonstration of one or more of these microorganisms in material obtained from the upper respiratory passages in life, though of value, furnishes us no definite evidence that the organism which has been identified has entered the lung and passed from the bronchi to produce pneumonia.

Study of autopsies following examination of the sputum during life has shown that an individual primarily attacked by influenza may suffer with a succession of pneumonias, one microorganism having prepared the way for another. The complexity of the subject is much increased by the truth that pyogenic microorganisms, like the tubercle bacilli, are capable of producing a considerable variety of pulmonary lesions.

Examination of the lungs of a large number of individuals who have died as the result of pneumonia following influenza has disclosed a succession of acute and chronic diseases. Immediately succeeding the height of the epidemic of influenza, death occurred with acute lobar pneumonia or with diffusely distributed hemorrhagic bronchopneumonia caused in the majority of instances by Pneumococcus IV in association with B. influenzæ. Superimposed infection with hemolytic streptococci increased in frequency and in individuals who had occupied certain wards was almost invariable. At a later period, from one to two months following the maximum incidence of influenza chronic lesions, namely, bronchiectasis, unresolved pneumonia, and chronic empyema were common and often occurred as the result of influenza which had had its onset at the height of the epidemic.

When influenza attacked the encampment, about 50,000 troops were quartered in it, and for a considerable period no more troops were brought into the camp and none left 109it. All cases of pneumonia occurring among these troops were brought to the base hospital so that the autopsies which were studied were representative of all the pneumonias following influenza in this limited group of men. It is noteworthy that autopsy disclosed no instance of fatal influenza unaccompanied by pneumonia.

Pneumonia of Influenza.—Knowledge concerning the bacteriology of the pneumonia of influenza dates from the study of the epidemic of 1889–90. The frequency with which Diplococcus lanceolatus occurred in association with influenzal pneumonia was well recognized, although several observers, notably Finkler^[57] and Ribbert,^[58] found Streptococcus pyogenes so often that they attributed the pneumonia of influenza to this microorganism.

During a subsidiary outbreak of influenza occurring in 1891–92 Pfeiffer^[59] discovered the microorganism which he believed was the cause of the disease. Pneumonia, he believed, was caused by the invasion of this microorganism into the lung, and the pneumonia of influenza, if death occurred at the height of the disease, was characterized not only by the presence of the bacillus of influenza, but was recognizable by its anatomic peculiarities. He described lobular patches of consolidation which were separated from one another by air containing tissue or were confluent, so that, although the lobular character was still recognizable, whole lobes might be affected. The consolidated tissue was dark red and within each lobular area were small, yellowish gray spots varying in size from that of a pinhead to a pea. In the mucus of the larynx and trachea were numerous microorganisms, including diplococci and streptococci, among which influenza bacilli were predominant; in the larger bronchi, bacteria other than influenza bacilli were less abundant, whereas in the finer bronchi filled with 110purulent fluid and in the lung tissue influenza bacilli had undivided sway. Pfeiffer stated that the changes described were found when death occurred at the height of the disease, whereas other pulmonary lesions might be sequelæ of this typical influenzal pneumonia.

Observations upon the pathology of influenzal pneumonia made during the epidemic of 1889–92 have been collected in the monograph of Leichtenstern^[60] published in 1896. He combats the opinion held by some observers that pneumonia with influenza is always catarrhal and cites many writers to prove that lobar pneumonia not infrequently accompanies the disease. Indeed, some have found “croupous” pneumonia more often than “catarrhal.” Krannhals^[61] at Riga found typical fibrinous pneumonia in 53 instances, doubtful forms in 22 and bronchopneumonia in 37. Cruickshank^[62] in England found croupous forms predominant. Among 43 autopsies performed upon individuals dead with influenza Birch-Hirschfeld^[63] found 11 instances of croupous lobar pneumonia, 8 instances of croupous lobular pneumonia and 24 instances of catarrhal pneumonia. Leichtenstern thinks that the atypical symptomatology of lobar pneumonia with influenza—for example, the purulent character of the sputum—has led many physicians to believe that lobar pneumonia rarely occurs. It is equally true that many instances of confluent lobular pneumonia are mistaken for lobar.

There appears to be no comprehensive description of the pneumonias of influenza based upon the epidemics of 1889–92. Descriptions dating from this period are much obscured by attempts to separate croupous or fibrinous from catarrhal pneumonias. Croupous lobular pneumonia has been recognized, for example, by Birch-Hirschfeld. Leichtenstern describes a form of pneumonia which he regards as 111neither lobar nor lobular although it implicates whole lobes; the consolidated tissue is homogeneous and varies in color from fleshy red to bluish red; it is tough and elastic in consistency. The author thinks that it is an error to regard this lesion as a confluent lobular pneumonia.

Kuskow^[64], who has discussed the pathology of influenza in considerable detail, has seldom seen lobar pneumonia but has almost invariably found, even when there is lobar distribution of the lesion, lobular patches of consolidation involving groups of lobules, single lobules or only parts of lobules; the lung tissue has been hyperemic and in places edematous.

Opinions concerning the pathology and bacteriology of the pneumonias of influenza, published since the recent epidemic, have varied almost as much as those just cited. Few observers have had the opportunity of making a considerable number of observations under conditions which determine the relation of the pulmonary lesions to the primary disease.

Keegan^[65] has found with influenza a massive and confluent bronchopneumonia, frequently resembling lobar pneumonia but distinguishable particularly in its early stages when the cut section of the consolidated lung has a finely granular character and each bronchiole stands out as a grayish area with intervening dark red alveolar tissue.

Symmers^[66] states that the pneumonia of influenza is a confluent lobular exudative and hemorrhagic pneumonia which bears a close resemblance to the pneumonic variety of bubonic plague.

Our commission^[67] published a preliminary report on pneumonia following influenza observed at Camp Pike. The occurrence of purulent bronchitis, distention of lungs, 112peribronchial hemorrhage and bronchiectasis were described. B. influenzæ was isolated from the bronchi in approximately 85 per cent of instances of influenzal pneumonia but from the consolidated lung in less than half this number of autopsies. Lobar pneumonia was present in a large proportion of autopsies but was less frequent than bronchopneumonia. Bronchopneumonic consolidation is in part red, lobular and confluent, in part nodular; pneumococci have a predominant part in the production of these lesions. Three types of suppurative pneumonia are described: (a) Abscess caused by hemolytic streptococci usually in contact with the pleura and accompanied by empyema; (b) Suppuration of interstitial tissue of the lung caused by hemolytic streptococci and accompanied by empyema; (c) multiple foci of suppuration clustered about a bronchus of medium size and caused by staphylococci. We have expressed the opinion that B. influenzæ descends into the bronchi; pneumococci, usually Type IV, may enter the lung and produce either lobar or bronchopneumonia. Hemolytic streptococci may descend and infect the pneumonic lung causing death often before suppuration has occurred. Epidemics of such superimposed streptococcus pneumonia in wards of the hospital were described.

LeCount^[68] says: “The pneumonia of influenza is commonly referred to as bronchopneumonia. It may be so designated, but it differs from other bronchopneumonias in its predilection for the periphery of the lungs and in the extent to which the inflammation is hemorrhagic.”

MacCallum^[69] states that the following types of pneumonia following influenza may be recognized. 1. Pneumococcus pneumonia. The lobular character of the consolidation is in these cases well marked, although it tends to lose its definiteness through the confluence of adjacent areas. The cut surface of the lung shows in the more acute cases a 113peculiar lobular or confluent consolidation which corresponds well with what is commonly written of the stage of engorgement in the description of lobar pneumonia. Later stages in the pneumonia show within these areas patches of rough gray consolidated tissue from which definite plugs of exudate project. 2. Staphylococcal pneumonia. 3. Streptococcal pneumonia. There is lobular pneumonia, the interlobular septa are edematous and, microscopically, bronchi and alveoli are loaded with streptococci. Whole areas of lung, though retaining their form, are entirely necrotic. Lymphatics are distended with exudate containing streptococci in great numbers, but in none of these cases is interstitial bronchopneumonia found. 4. Pneumonia produced by B. influenzæ of Pfeiffer. The lung is studded with palpable shot-like grayish yellow nodules; the bronchi exude thick yellow pus, which contains influenza bacilli. Microscopically, the walls of the bronchi are found to be thickened by mononuclear infiltration and new formation of connective tissue. The walls of the alveoli are thickened and indurated and the alveoli often contain fibrin in process of organization. Absence of conspicuous changes in lymphatics, absence of intense pleural infection and relatively scant numbers of polynuclear leucocytes in the exudate within the alveoli and bronchial walls are, MacCallum states, all that distinguish this pulmonary change from the interstitial bronchopneumonia caused by the hemolytic streptococcus and described by him in previous papers.

Lyon^[70] designates the pulmonary lesion following influenza, hemorrhagic pneumonitis, the lung tissue containing serous fluid loaded with red blood corpuscles; in many instances there was such scant consolidation that the process could not be regarded as a true pneumonia. In 35 instances the lesion was lobular in distribution and in 16 114instances was sufficiently extensive to be designated lobar, but it was not typical lobar pneumonia, and, often associated with lobular patches of consolidation, appeared to be a confluent lobular pneumonia.

Goodpasture and Burnett^[71] say: “The difficulties of analyzing the pulmonary lesions in any group of influenza pneumonias as they have appeared in this epidemic, are very apparent to anyone who has had an opportunity to observe the bacteriology and pathology of this accompaniment of the disease.” There is an acute outflow of the fluid elements of the blood and of hemorrhage into the lung tissue filling the alveoli in lobular areas and not infrequently in an entire lobe. By a special method of staining these authors have studied the distribution of Gram-negative bacilli with the morphology of B. influenzæ. The fact that in certain very early cases demonstrable bacteria of any kind are scarce or not found at all, has lead them to believe “notwithstanding the demonstration of influenza bacilli in pure culture in the lung in all but one instance, that at this stage organisms are comparatively few within the alveoli, and the primary injury is due to a very potent toxic agent elaborated in and disseminated through the larger air passages. In the later stages or from the beginning, if the injury be slight, the infection focalizes about the bronchi or their terminations, so that the bronchial and lobular distribution becomes very conspicuous.” Typical lobar pneumonia with “croupous” exudate within the alveoli occurs in cases complicated by secondary pneumococcus infection.

Wolbach^[72] has found that two types of pneumonia are characteristic of influenza. In cases in which death has occurred within a few days after onset of pulmonary signs, the lung tissue is dark red and “meaty in consistency” and contains abundant blood-tinged serous fluid which drips 115from the cut surface. The other type of lesion is found in patients who have lived for ten days or more after onset of the disease; there is extensive bronchitis, bronchopneumonia, discrete or confluent, and peribronchitis. The lungs are voluminous and the smaller bronchi are distended. Microscopically, there is peribronchitis with extensive infiltration of the interlobular septa and organization in alveoli and bronchioles. This lesion is that designated by MacCallum as “interstitial bronchopneumonia.” Wolbach thinks that the two types of lesion represent different stages of the same process. He regards as distinctive of the pneumonias of influenza the presence of hyalin fibrin lining distended air spaces. With the two types of lesion which have been described, B. influenzæ was the only organism which could be cultivated, and the author associates these distinctive conditions with that microorganism.

Streptococcus Pneumonia.—Finkler emphasized the importance of streptococcus pneumonia as a complication of influenza. In 1888 he described instances of acute primary streptococcus pneumonia observed in 1887 and 1889. This form of pneumonia, Finkler thought, occurred in Bonn in epidemic form before the influenza epidemic of 1889–90 and, he states, exhibited an astonishing similarity to the pneumonia of influenza. He thought that later there was in Bonn a mixed infection of influenza and primary streptococcus pneumonia. In one type of streptococcus pneumonia, described by Finkler, there was lobular consolidation which in multiple patches produced “pseudolobar” consolidation; the consolidated lung was smooth and red, and similar to spleen, rather than hepatized. In another group of instances the lesion merited the name “erysipelas of the lung.” The lesion was an acute interstitial pneumonia in some places, a cellular or occasionally fibrinous pneumonia with involvement of the interstitial tissue in other places. There was edematous swelling and accumulation of leucocytes in the interstices between the alveoli and about 116the blood vessels and bronchi. Finkler stated that the similarity to erysipelas might suggest that the lymphatics contain streptococci, but this relation did not exist although large lymphatic channels were occasionally filled with coagulum. He asserted that the disease was contagious and cited cases which he believed had their origin in hospital wards.

The widespread occurrence of streptococcus pneumonia in the army camps in this country attracted attention during the first months of 1918. Cummings, Spruit and Lynch^[73] at Fort Sam Houston, Texas, recognized the prevalence of streptococcus pneumonia, both as a complication of measles and in association with lobar pneumonia, and showed that the microorganism concerned was a hemolytic streptococcus. In 7 autopsies upon individuals with lobar pneumonia, they found pneumococcus alone in 2 instances and pneumococcus and hemolytic streptococcus or hemolytic streptococcus alone in 5 instances. Hemolytic streptococcus was found in all of 24 instances of bronchopneumonia, three-fourths of which followed measles. They recommend the bacteriologic examination of the throat of patients with measles and the segregation of those who harbor hemolytic streptococci.

Cole and MacCallum^[74] have published almost simultaneously, with that just cited, a report upon the pneumonia which has occurred at Fort Sam Houston and have shown the importance of hemolytic streptococci in its causation. They have found two varieties of pneumonia, namely, acute lobar pneumonia which does not differ essentially from that which occurs elsewhere and bronchopneumonia which in most cases has followed measles and is caused by S. hemolyticus. They think this infection is usually acquired in the hospital. They believe that the pulmonary lesions are characteristic. In this publication and elsewhere MacCallum 117has designated the lesion “interstitial bronchopneumonia.”

The epidemic of streptococcus pneumonia and empyema occurring at Camp Dodge, Iowa, from March 20 to May 10 is described by Miller and Lusk^[75]. During this period there were 400 cases of pneumonia, whereas from September 20, 1917, to March 20 there had been only 276 instances of lobar pneumonia. The type of pneumonia changed, there was more severe intoxication and empyema became very frequent; in 85 of 95 exudates streptococci were found. The outbreak of pneumonia bore no relation to measles. The authors state that a mild tracheitis was prevalent in the cantonment during March, and whenever a large group of soldiers congregated coughing was noticeable.

MacCallum^[76] studied the pneumonia at Camp Dodge during May and found 17 instances of the lesion which he had designated interstitial bronchopneumonia; of these, 9 followed measles, although in the earlier part of the epidemic there appear to have been, he states, few such cases. Cultures made at autopsy, except in a few fatal cases of uncomplicated lobar pneumonia caused by the pneumococcus, showed the hemolytic streptococcus in every situation throughout the respiratory tract and pleura.

The pneumonia which occurred at Camp Funston is of special interest to our commission because we were for a time stationed at this camp and had the opportunity of following in the excellent records of the hospital the history of the occurrence of pneumonia during the year following the establishment of the camp in September, 1917. Stone, Phillips and Bliss^[77] have described the outbreak of pneumonia which occurred in March, 1918. At this time there was, the authors state, severe pneumonia with frequent empyemas due to hemolytic streptococci. This condition 118which did not follow measles was responsible for the greatly increased death rate in March; 9 deaths occurred in February, 45 in March, 25 in April and 14 in May. They found during March 26 instances of multiple pulmonary abscess. In 29 autopsies they found a pleural lesion which they designate “subcostosternal pus pockets”; it occurs only in association with empyema caused by hemolytic streptococci.

Our commission^[78] has shown that an epidemic of influenza, well characterized by its epidemiology and symptoms, preceded and accompanied the outbreak of pneumonia just described. Between March 4 and 29 1,127 men from Camp Funston, which then contained 29,000 men, were sent to the base hospital with influenza and many more were treated in the infirmaries of the camp; on March 11 107 patients with influenza were admitted to the hospital. The greatest incidence of pneumonia in the history of the encampment up to this time occurred between March 9 and 29, immediately following the outbreak of influenza, the maximum incidence of pneumonia occurring five days after the maximum for influenza.

The foregoing observations are cited to prove that streptococcus pneumonia, which occurred during the spring of 1918 at Camp Funston and doubtless at other camps, had its origin in influenza and did not differ in character from that which occurred on a much larger scale in the fall of 1918.

Table of Autopsies.—In order to present as briefly as possible the data upon which the present study has been based, autopsies have been assembled in tabular form in the order of their performance (Table XXVII). During the early period of the epidemic autopsies were performed on all who died with pneumonia, but later, with increase in the number of deaths, this became impossible and autopsies were performed on all those who died in certain wards.

119Comparison of charts representing incidence of influenza and of deaths from pneumonia furnishes evidence that fatal pneumonia during the period of investigation was with few exceptions referable to influenza. During two weeks, namely, from September 1 to 14, before the presence of the epidemic was evident, there were only 2 fatal cases of pneumonia. In most instances the relation of pneumonia to influenza is established by a definite history of influenza having its onset during the epidemic. Bronchopneumonia usually develops gradually as a sequence of influenza in which purulent bronchitis has occurred. Lobar pneumonia may develop in cases of influenza complicated by purulent bronchitis. In some instances there is apparent recovery from influenza indicated by return of temperature to normal; after from one to three days of normal temperature there is typical lobar pneumonia with rusty sputum. In many instances of pneumonia having their onset at the height of the epidemic of influenza, the history indicates that pneumonia was present immediately after the onset of symptoms, so that the onset resembled that of pneumonia rather than of influenza.

Cases of pneumonia following measles have been excluded from the table in order that they may be studied separately and compared with the pneumonias of influenza. It is noteworthy that the lesions of pneumonia following measles have shown a very close resemblance to the pneumonias of influenza, with regard both to pathologic characters and to bacteriology.

Five instances of pneumonia following typhoid fever (Autopsies 245 and 329), scarlet fever (Autopsy 311) or mumps (Autopsies 403 and 417) have been excluded from the table. These secondary pneumonias are grouped as an appendix to the section on pneumonia following measles. It is not improbable that individuals with the diseases named are just as susceptible as others to influenza. Included in the table is an instance (Autopsy 487) in which a definite attack of influenza preceded scarlet fever.