ERRORS IN THE DIAGNOSIS OF 
PULMONARY TUBERCULOSIS 
JOSEPH WALSH, M.D. 
PHILADELPHIA ERRORS IN THE DIAGNOSIS OF 
PULMONARY TUBERCULOSIS * 
JOSEPH WALSH, M.D. 
PHILADELPHIA 
My experience in connection with the diag- 
nosis of tuberculosis has been one of surprise, 
not at the number of errors made, but at the ac- 
curacy of the conclusion when care was exer- 
cised in the examination. In the 471 necropsies 
at the Phipps Institute there were six cases 
which were not tuberculosis; of these six two 
were properly diagnosed (both cardiovascular 
disease) and four were diagnosed tuberculosis, 
though one was carcinoma, one actinomycosis 
and two cardiovascular disease. This represents 
an error of less than 1 per cent. These sta- 
tistics constitute my personal experience with 
advanced cases diagnosed by about twenty dif- 
ferent physicians and, coming as they do from 
necropsies on consecutive deaths, represent, I 
think, what a study of the physical signs can 
accomplish. 
It is more difficult to determine the errors in 
early cases, since they do not come to necropsy, 
yet not so difficult as might be imagined on first 
thought, because subsequent developments, like 
the later discovery of tubercle bacilli in the 
sputum, or death from another cause with 
necropsy, often come to our aid. Again, my 
experience has been one of surprise at the ac- 
* Read before the West Branch of Philadelphia County 
Medical Society, Jan. 20, 1920. 4 
curacy of the diagnosis when the cases were ex- 
amined with care. 
The errors I have encountered have prac- 
tically always been the result of haste and negli- 
gence, like the examination of one part but not 
all parts of the body, the failure to examine the 
sputum bacteriologically, the failure to try all 
the physical signs, or to correlate signs and 
history. 
The first error I ever saw was when I was 
a resident physician, and was due to failure to 
examine the body generally. An old man had 
been admitted for alcoholism in a pretty 
drunken state. The examination was as limited 
as such cases frequently experience. When 
about to be discharged a running sore was 
noticed on the side of his neck, which was 
hastily diagnosed tuberculosis of the cervical 
glands, and operation advised. The operation 
was not a success, since, to our surprise, the 
bottom of the suppurating tract extended below 
the suprasternal notch, and could not be reached 
with the curet. The patient died about ten days 
later, and necropsy showed a cancer of the pan- 
creas as large as two fists, which had infiltrated 
behind the esophagus to the neck where it had 
opened. The most casual examination of the 
abdomen during life would surely have revealed 
the tumor. 
It is remarkable the typically advanced cases 
which fail to show tubercle bacilli in the sputum 
on a number of examinations during life, and 
yet show them in every part of the lung after 
death, and everyone has gradually come to rec- 
ognize it and not allow too much weight to the 5 
negative finding. On this account, we are some- 
times negligent in the examination of the spu- 
tum of an early case, thinking even though it is 
tuberculosis, it will not show tubercle bacilli. 
As a consequence I have seen several errors like 
the following: 
A man of 45, thirty pounds over the average weight 
for his height and age, presented himself with a cor- 
rect diagnosis of diabetes. He had 5 per cent, of sugar. 
He had also, however, a history which had been over- 
looked, that three brothers had died of tuberculosis, 
and that he had coughed to a slight extent off and on 
for a number of years, and had always expectorated. 
Like lay people frequently, he thought this small 
amount of cough and expectoration could be normal. 
•Examination of the chest on six different occasions 
has failed to reveal any abnormality by inspection, 
palpation, percussion or auscultation, except a few fine 
moist rales over the left upper lobe on cough. He is 
splendidly developed, inclined to be stout, and as far as 
his personal feelings are concerned is perfectly well. 
His sputum shows tubercle bacilli. In other words, in 
addition to his diabetes, he has at least a mild tuber- 
culosis of the lungs. 
Another case illustrating the same is a student, 
aged 21, with a father, mother and eight brothers and 
sisters well. He came in with a diagnosis of essential 
asthma lasting over thirteen years. He had very little 
cough and expectoration, so little that he scarcely 
noticed either. The physical examination revealed 
only impairment at the right apex with asthmatic rales 
over the chest; the expectoration, though very small 
in amount and hard to procure, showed such a number 
of tubercle bacilli that the specimen was preserved for 
exhibition purposes. 
Though the failure to find tubercle bacilli in 
the sputum does not negative the diagnosis, their 
presence is so absolutely indicative that care 
must be exercised to avoid the possibility of 6 
confusion of specimens. Many years ago I did 
a necropsy on a case in West Philadelphia with 
Dr. Bryan: 
An old woman of 75 had been a patient in a hospital 
for the insane for a number of years. She had a slight 
hacking cough and had gradually emaciated. The 
diagnosis in the hospital was that of the form of her 
insanity and tuberculosis of the lungs. The chart not 
only stated that there were tubercle bacilli in the 
sputum, but gave an actual diagram of a microscopic 
field showing about a dozen scattered microorganisms. 
Externally, she showed the typical appearance of ad- 
vanced tuberculosis, being emaciated to a skeleton, so 
that with this diagnosis in front of us we had no 
doubt what we were going to find; yet, the necropsy 
showed perfectly normal lungs and larynx, and we 
were obliged to conclude that the finding of tubercle 
bacilli in the sputum was due to a mixing of speci- 
mens and that her emaciation was probably the result 
of the unconscious and unnoticed starvation sometimes 
seen among the insane. An interesting feature of this 
case was that the patient's daughter of 20 was at the 
time suffering from chlorosis, and in the light of the 
mother's tuberculosis, the question was seriously con- 
sidered as to whether it was essential chlorosis or only 
the chloranemia of tuberculosis. The necropsy cleared 
the atmosphere. 
The failure to find tubercle bacilli in the 
sputum is much more common, though a com- 
plete physical examination rarely leaves one in 
doubt: 
On account of continuous, even though slight, cough 
and expectoration a woman had her sputum examined 
through her own initiative in a commercial laboratory 
once a year for ten years with negative findings. 
Though she urged her physician frequently to go over 
her chest he usually refused, because she had no family 
history of tuberculosis, and the general condition ap- 
peared good. Rarely when he did examine it, the 7 
examination was so superficial, being made through the 
clothes and limited to two or three signs at the tops 
of the lungs, that she herself doubted its conclusive- 
ness, yet she never went any further. At the end 
of the ten years she had an acute manifestation, and 
examination revealed chronic tuberculosis of the whole 
of the right upper lobe, and acute tuberculosis of most 
of the left lung. Tubercle bacilli were now found 
in abundance. 
This failure to make a real physical exami- 
nation is responsible for practically all the 
errors, and I have notes on twenty somewhat 
similar cases of tuberculosis in which examina- 
tion of the chest was neglected. The following 
were the diagnoses under which they were being 
treated: Malaria, 5; indigestion, 4; chronic 
rheumatism, 3; lumbago, 1; neurasthenia, 2; 
chronic bronchitis, 2; overwork, 1; heart dis- 
ease, 1; arteriosclerosis, 1. 
The explanation of this neglect appears to be 
that the medical journals dilate on the impor- 
tance of a particular sign or the frequency with 
which tuberculosis begins in a particular place, 
and physicians in haste proceed to limit them- 
selves to that particular sign or place. 
Several papers have appeared recently in the 
United States on the precision with which the 
whispered voice sounds outline a consolidated 
area either pneumonic or tuberculous. No one 
employing the stethoscope frequently can have 
failed to notice this phenomenon, and more than 
one of us has thought when he first remarked it, 
that he was on the probable trail of a discovery. 
Further study, however, has shown its decep- 
tiveness, for though it is surprising the number 
of cases in which* the whispered voice sounds 8 
outline a tuberculous lesion better than any 
other sign, the number of cases in which it fails 
is so large that we were prevented from dis- 
cussing it, especially since the general fact had 
already been called attention to by both Laennec 
and Walshe before 1850. 
The French have recently been writing on 
what Chauvet calls the zone d'alartne, the zone 
of suspicion in pulmonary tuberculosis, the area 
in which the earliest signs of tuberculosis are 
most frequently manifested. This area lies be- 
tween the first and third dorsal vertebrae, and 
alongside the vertebral column. If a 25 cent 
piece is glued to this area with a little collodion 
and the lungs roentgen-rayed, it will be found 
to correspond with the summit of the lung more 
nearly than the more external supraspinous 
fossa which is so frequently studied. 
Kronig's isthmus has had more than its share 
of attention, for though it aids in doubtful cases, 
it is only an aid, and nothing like a finished con- 
clusion should be based on it. Personally, I 
have never seen anyone able to go over the chest 
with one sign like percussion or whispering 
pectoriloquy and come to a conclusion, or em- 
ploy all the signs over one part, like an apex, 
and describe its condition. 
As a refinement, knowledge of this character 
is worth while, provided we do not make par- 
ticularly a negative diagnosis dependent on it, 
and pay no attention to other parts or signs. 
The Phipps Institute experience proves how ac- 
curately the physical signs locate and describe a 
lesion provided all of them are employed over 
every part of the chest, yet we find it just as re- 9 
markable how one or two signs may be entirely 
negative, or even at variance with the expected. 
We sometimes found, for instance, a normal 
percussion note over a pneumonia, but ausculta- 
tion then indicated the consolidation,- or again 
negative auscultatory signs over a massive fibro- 
sis, but the percussion note set us straight. 
About six years ago it was promulgated that 
headache or localized pains elsewhere in the 
body associated with slight fever and general 
symptoms pointing towards infection might be 
the result of an unopen pocket of pus at the root 
of a tooth. The number of teeth of poor tuber- 
culous patients which have been uselessly ex- 
tracted since that time is uncountable. A man 
came in three years ago with far advanced tu- 
berculosis of both lungs, a typical history of 
gradual advance of the disease, and no teeth, all 
having been extracted in groups of three or 
four with the idea that the symptoms would be 
relieved, and though the symptoms pointed even 
then to the lungs, the lungs were never ex- 
amined. Since then I have seen five somewhat 
similar cases. 
Before the time of Laennec the diagnosis of 
tuberculosis of the lungs was made on the his- 
tory, the symptoms and the general appearance; 
Laennec's work promulgated our modern physi- 
cal signs and produced the possibility of earlier 
diagnosis, but did not take away the necessity 
for the study of the history and the symptoms. 
Koch's discovery made the diagnosis so fre- 
quently absolute at an earlier stage that we 
learned how to recognize the disease at this 
earlier stage even apart from the finding of 10 
tubercle bacilli, but this is possible only by a 
correlation of all of our previous knowledge. 
Tuberculin once more helped to a still earlier 
diagnosis, but acts only as an aid to what we can 
learn apart from it. Now we have a new ad- 
dition, the roentgen ray, but like everything else 
so far produced, it is only a help to the knowl- 
edge acquired by other means. Yet it is follow- 
ing in the footsteps of the previous discoveries, 
and is being used by some as the sole arbiter, 
as were in their day the physical signs, the pres- 
ence or absence of tubercle bacilli, or the tuber- 
culin reaction. That it cannot be depended on 
alone is evident from numerous examples we 
are constantly encountering, like the following: 
Girl of twenty went to the family physician com- 
plaining of cough, expectoration, and loss of ninety 
pounds in weight. Physical examination was neglected 
on account of the convenience of the roentgen ray. 
The roentgen ray showed shadows in the right chest. 
On these shadows, the extreme loss of weight, the 
cough and expectoration, the diagnosis of tuberculosis 
of the right lung was made. On the plate the heart 
appeared normal in size. Physical examination showed 
the heart apex in the sixth interspace and anterior 
axillary line, and the right border three-fourths of an 
inch to the right of its normal situation, with a slight 
presystolic murmur, and a blood pressure of 240 
systolic and 170 diastolic. There were no signs in the 
lungs except prolonged expiration everywhere. About 
three weeks after she developed cerebral apoplexy with 
right-sided paralysis and died two weeks later. 
A boy of eight with pleuropneumonia last March 
followed by continuation of cough and expectoration 
was treated for tuberculosis until November 10. Ex- 
amination showed flatness from the second rib and 
fourth dorsal spine down on the right with the apex 
of the heart outside the midclavicular line, and the 11 
liver slightly lower than normal. Above the flatness 
there were the typical skodaic signs of tympany, am- 
phoric breathing and whispering pectoriloquy, usually 
found above an empyema or an effusion, and since the 
temperature curve was that of empyema, a diagnosis of 
empyema was made. A roentgen-ray examination 
showed the upper lobe normal with a dense shadow 
over the upper part of the lower lobe, and a clear 
space in the lower outer margin in such a way as to 
make it appear that there was normal lung below the 
pathologic condition. The roentgenologist suggested 
that the condition was more likely a pulmonary abscess 
than an empyema, or if an empyema, it was practically 
surely loculated. Operation showed the lung pressed 
up against the spinal column throughout its entire 
lower two thirds, with pus in the pleural cavity extend- 
ing to the base; in other words, the physical signs 
were correct and the roentgen-ray diagnosis wrong, 
though the reason for the roentgen-ray error was 
not discovered. 
In General Hospital No. 17 we had necropsies 
on six cases in which very careful physical and 
roentgen-ray examinations had been made. In 
these comparisons 1 it was a pleasure to see how 
well both stood up, especially in the diagnosis 
of the general condition, and though we found 
that the physical signs could be more uniformly 
depended on, we also found that in obscure 
cases the roentgen ray proved a valuable aid. 
Some of the roentgen-ray errors were as 
follows: 
Case No. 197. Physical signs and necropsy both 
showed a cavity at the apex of the lower lobe sur- 
rounded by caseation with scattered tubercles be- 
low ; the roentgen ray showed a dense shadow over 
this whole area making the scattered tubercles look 
like massive caseation and gave no indication of cavity. 
Case No. 395. Physical signs and necropsy both 12 
showed compensatory emphysema; the roentgen ray, a 
large cavity in the center of the lobe with fibrosis at 
the apex. 
Case No. 439. Physical signs and necropsy both 
showed massive caseation at the apex with scattered 
tubercles below; the roentgen ray, a normal condition. 
2026 Chestnut Street. 
Heprin Ied f rom the Pennsylvania Medical Journal 
March, 1920, Vol. XXIII, p. 333