THE RESPONSIBILITY OF THE TONSIL IN TUBERCULOUS ADENITIS. BY FRANK S. MATHEWS, M.D., OF NEW YORK, Associate Surgeon to St. Mary's Hospital for Children; Assistant Surgeon to the General Memorial an4 St. Francis Hospitals. To attempt to prove that the gland at the angle of the jaw may be infected with tuberculosis by way of the tonsil with which it is in such immediate anatomical and physiological relation would seem a demonstration of the obvious. No one is likely to dispute it, but when we ask whether the tonsil is the uniform, the occasional, or a rare source of cervical ade- nitis, we may expect widely different opinions. For its frequent etiological relationship we have the fol- lowing: Pathological conditions are frequent in children in both tonsils and lymph-nodes. When the tonsil is inflamed from any cause the tonsillar nodes at the angle of the jaw are the first to enlarge. One familiar with the early stages of tuberculous adenitis in children will agree that in at least go per cent. of cases the first nodes to enlarge are these same tonsillar nodes. Lymphatic anastomosis being very free, this does not rule out the possibility of infection of these nodes from a back tooth, the pharynx or the nasopharyngeal ade- noids. Later in life infection of both tonsil and lymph-nodes is less common. Wood, of Philadelphia, says on the basis of personal observations, that the tonsils of consumptives show evidence of tuberculosis in the vast majority of cases, though the rest of mouth, nose, and pharynx does not. Dr. A. M. Shrady, with a very long experience at the Seton Hospital for Tuberculosis says it is a rare occurrence for consumptives to develop clinical tuberculosis of the tonsil aird that they usually die without clinical signs of infection of neck nodes. On the other hand the infrequency of clinical tonsillar tuberculosis and hitherto the almost uniform failure to demonstrate it micro- scopically have made some question the etiological importance - 27 753 C5C) 754 FRANK S. MATHEWS. of the 'tonsil in cervical tuberculosis. Other avenues of in- fection are by no means wanting, such as gums, nose, and pharynx, and even the skin of the face, lips, and scalp. In the past ten years we have seen probably 25 cases of tubercu- lous femoral adenitis result from scratches on the foot. One child under six months of age had a tuberculous inguinal bubo following a scratch with a diaper pin at the base of the penis. These cases are mentioned simply to show that almost any part of the body's surface may admit the bacillus. The writer submits the results of the examination of sixty-five whole tonsils recently removed. These. can be considered in three groups: i. Fifty-seven tonsils removed for a variety of reasons from children and embracing all types of tonsils. In none of these patients was there reason to suspect tuberculosis in the neck or other part of the body. None showed tuberculosis on microscopic examination in the tonsil, i.e., there was no acci- dental discovery of tubercle in a tonsil where it was not sus- pected. This confirms the experience of others. Hodenpyl examined several hundred tonsils without finding any tuber- culous. Hurd quotes Wright as having examined 6o unse- lected specimens without finding tubercle. Judd, of Rochester, says that at the Mayo's Clinic only one or two tuberculous tonsils have been found among some hundreds of tonsils sec- tioned. In conversation one frequently hears it said as a piece of common knowledge that tonsils removed for all causes are found tuberculous in 5 to IO per cent. of cases. I am unable to find who said it first and feel certain that it does not approxi- mate the truth. With the uniform result of examinations by Hodenpyl, Wright, the Rochester Clinic and myself we feel safe in asserting that one is not likely to find the tonsil harboring the tubercle bacillus without other manifestations of tuberculosis in the body. 2. Five tonsils showing well-defined tuberculous lesions and all taken from children with recent tuberculous cervical adenitis. CASE I.-A small collection of nodes at angle of left side of jaw, largest hickory-nut sized. Left tonsil small. In it, several THE TONSIL IN TUBERCULOUS ADENITIS. 755 discrete scattered tubercles, few giant cells, just a suspicion of central necrosis. A couple of tubercles located in the capsule of the tonsil. CASE II.-Boy at Sea Breeze Hospital for Tuberculosis. Adenitis has not been operated on. Has widely distributed enlarged nodes in both sides of neck. Tonsils are peppered with discrete miliary tubercles both near mouth surface and capsule. No giant cells or necrosis. CASE III.-Age three years. A collar of enlarged nodes in both sides of neck. Seem recent. Von Pirquet positive. Ton- sil quite small. Small area in one tonsil shows giant cells with peripheral nuclei and poorly defined epithelioid tubercles. After removal of tonsils the glandular enlargement partly subsided and child removed from hospital without operation upon the neck. CASE IV.-Already reported in ANNALS OF SURGERY, April, I910, P. 577. The enlargement of neck nodes followed a ton- sillotomy. Discrete typical miliary tubercles found in tonsillar stumps. CASE V.-Child, age eight years. A few soft glands at angle of jaw found to be tuberculous. The tonsil on the same side as the nodes showed many typical miliary tubercles. The en- largement of nodes was recent. 3. Cases with tuberculous glands in neck but zwrithout tube.- culosis in tonsil. CASE VI.-Child, age three. Glands in right side of neck of one month duration. Von Pirquet positive. The largest node is the tonsillar one, but the first one to enlarge was in front of the submaxillary salivary gland and it was the only gland at oper- ation which had gone to the stage of suppuration. Tonsils, small, appear normal and no tuberculous lesions found in them. In this case it would seem likely from location of swelling that infection took place through lip, gum, or floor of the mouth. CASE VII.-Child. Tonsillar nodes enlarged. Also a large soft node in the substance of the parotid gland just in front of external auditory canal. Von Pirquet positive. Nodes proved tuberculous. Tonsil without tuberculosis. The clinical evidence would suggest that this was a case of scalp infection as the most advanced tuberculous lesion was in a node well up in the parotid gland. 756 FRANK S. MATHEWS. CASE VIII.-Two good-sized tonsils not found to be tuber- culous. Von Pirquet positive. A collection of nodes in right tonsillar group of several months' standing proved to be tuber- culous. In this case clinical evidence suggested tonsillar origin. As illustrating the probability with which we can infer the point of infection from the location of the first involved node, we may mention the following case: A child, aged four, had tuberculous nodes on the right side in the tonsillar and sub- maxillary region. A submaxillary node was suppurating. There was a red spot on the upper gum, at the insertion of the second incisor tooth, about one-fourth inch in diameter. This little piece of gum was removed at the same time as the nodes and showed typical tuberculous lesions. This is the only case in which we have demonstrated a tuberculous gingi- vitis as the starting point for tuberculous adenitis. In this case the tonsils appeared normal and were not removed. There is an impression that the tubercle bacillus may pene- trate the tonsil without producing any manifestations of its presence and be carried through it to the neck nodes. So far as I know, the only basis for the belief is one experiment by Wood of Philadelphia on a pig. He rubbed tubercle bacilli on the tonsil, killed the pig later and identified bacilli in the cervical glands only-not in the tonsil. To sum up, 57 tonsils of patients not clinically tuberculous show no tuberculous lesions. Of 8 patients with cervical adenitis the tonsils are tuberculous in five. In two there is evidence of other than tonsillar origin for the infection. In one the adenitis was of long standing and may have been extra- tonsillar in origin or slight tonsillar lesions may have escaped observation. These results closely parallel those of Hurd. In an ex- amination of ii tonsils from patients with tuberculous nodes he found tuberculosis in the tonsil 8 times. In an examina- tion of I3 tonsils of persons with slightly enlarged neck nodes, not believed from clinical evidence to be tuberculous, only one tonsil showed tubercle. In cases then where there is clinical evidence that the tonsil is responsible for a tuberculous adenitis, THE TONSIL IN TUBERCULOUS ADENITIS. 757 both Hurd and myself have as a rule found tuberculous lesions in the tonsil. CHARACTER OF TONSILLAR TUBERCULOSIS. Hurd has said that tuberculosis is to be found in the small buried tonsils, not the large soft elevated ones. We may put it in another way and say tuberculosis does not greatly enlarge the tonsil. In none of our five have the tonsils been consider- ably enlarged. The lesions as seen by the microscope have shown considerable variation in type. For instance, giant cells are in some few, in others abundant. In general we may say that the lesions are scattered rather than confluent, ulceration was never present and there was never more than a suggestion of central necrosis. In other words, we see the early or cellu- lar type of lesion rather than the late fibrous and necrotic. The tubercles have been found both superficial and deep- in one case even in the capsule. With a recognition of the frequency of tonsillar infection we shall probably more frequently remove the tonsil in cases of cervical adenitis than has been done in the past with the idea of avoiding reinfection of neck nodes. That we have been able to cure so large a per cent. of gland cases without remov- ing the tonsil is simply another illustration, if one be needed, that we do not cure surgical tuberculosis by removing every single bacillus and lesion but by reducing the load of infection with which the body must contend. If the tonsils and nodes are both to be removed, it would not seem wise to attempt both at one sitting. Which shall be done first will probably have to be decided on the merits of each case. We have seen the nodes largely subside on removal of the tonsils alone; in other cases the extensive involvement of -uch nodes makes their removal much more important than that of the tonsil. The material for this paper has mainly come from St. Mary's Hospital for Children. The pathological work has been done at the Laboratory of Surgical Pathology of Colum- bia University through the kindness of Dr. W. C. Clarke.