Volume 95.

Number 5 May 1988

THORACIC AND TheJournalof

CARDIOVASCULAR SURGERY
J THORAC CARDIOVASC SLRG 1988;95:747-57

Original Communications

"Maximal" thymectomy for myasthenia gravis

Results

Thymectomy has been shown to be effective in the treatment of myasthenia gravis. The logical goal of
operation has been complete removal of the thymus, but there has been controversy about the surgical
techniqueand its relation to results. Surgical-anatomic studies have shown gross and microscopic thymus
widely distributed in the neck and mediastinum. We believe that an en bloc transcervical-transstemal
"maximal" thymectomy is required to remove aU thymic tissue predictably. Ninety-five patients with
generalized myasthenia gravis underwent "maximal" thymectomy consecutively between 1977 and 1985
and were evaluated 6 months to 89 months after operation. In Group A (N = 72), myasthenia gravis
without thymoma, the uncorrected data revealed that 96% (69) had benefited from operation: 79% (57)
had no symptoms; 46 % (33) were in remission; 33 % (24) were symptom free when receiving minimal
doses of pyridostigmine; and none were worse. Life table analysis yielded a remission rate of 81 % at 89
months. In group B (N = 8), myasthenia gravis without thymoma for which patients underwent
reexploration for incapacitating weakness after earlier transcervical or transsternal operations, residual
thymus was found in all, One patient was in remission, two were symptom free when receiving medication,
one was unchanged, and none were worse. In group C (N 15), myasthenia gravis and thymoma, two
patients were in remission and nine were symptom free when receiving medication. Two patients in this
group died 2 and 4 years postoperatively in crisis. Response to thymectomy in group A was greater in
patients with mild myasthenia gravis and may have been better in patients who had symptoms for less than
60 months preoperatively, but the response did not depend on age, sex, presence or absence of thymic
hyperplasia or involution, or titers of acetylcholine receptor antibodies. The response to thymectomy in

Alfred Jaretzki Ill, MD, Audrey S. Penn, MD (by invitation), David S. Younger, MD
(by invitation), Marianne Wolff, MD (by invitation), Marcelo R. Olarte, MD (by invitation),
Robert E. Lovelace, MD (by invitation), and Lewis P. Rowland, MD (by invitation),
New York, N. Y.

From the Departments of Surgery, Neurology, and Pathology,
Columbia-Presbyterian Medical Center, New York, N.Y.
Supported by grants from the Muscular Dystrophy Association and
the National Institutes of Health (N.S. 17904).
Read at the Sixty-seventh Annual Meeting of The American Associ-
ation for Thoracic Surgery, Chicago, 11I., April 6-8, 1987.
Address for reprints: Alfred Jaretzki III, MD, Columbia-Presbyterian
Medical Center, 161 Fort Washington Ave.. New York. NY
10032.

748 Jaretzki et al.
group B was striking but slower than in group A, perhaps because symptoms were more severe and of
longer duration. The response in group C was also less good than in group A and proportionately fewer
benefited. These results support the recommendation for thymectomy in the treatment of patients with
generalized myasthenia gravis and indicate the desirability of a maximal procedure. For persistent or
recurrent severe symptoms after previous transcervical or submaximal traossternal resections, reoperation
by this technique is also recommended.
Myasthenia gravis (MG) is an autoimmune disease
in which the thymus gland is thought to playa central
role in the pathogenesis.' Weakness results from a
reduction in acetylcholine receptors (AchR) at the
neuromuscular junction.' MG is still a potentially debil-
itating and life-threatening illness despite advances in
therapy. Thymectomy is now standard therapy, and
improvements in surgical techniques, anesthesia, and
respiratory care and the use of plasmapheresis have
drastically reduced the operative morbidity. The mortal-
ity is close to nil regardless of the surgical approach.
There is, however, controversy about the surgical tech-
nique and its relation to results.
In 1977, preliminary surgical-anatomic studies indi-
cated that thymic tissue was widely distributed in the
neck and anterior mediastinum.' As a result of that
experience, and because of reports of microscopic foci of
thymus in pericapsular mediastinal fat," we5 developed a
combined transcervical-transsternal en bloc resection
that, we believe, comes as close to complete removal of
thymic tissue as is surgically possible. The term maxi-
mal rather than total thymectomy has been used
because microscopic foci or distant ectopic tissue may
remam.
Patients and methods
Patient selection. Between June 1977 and Jan. I, 1985, we
performed 124 consecutive "maximal" thymectomies. No
selection was made for age or sex or for duration or severity of
symptoms. Of the 124 patients, 95 were evaluated 6 to 89
months after operation. Primary operations were done in 72
patients for MG without thymoma (group A), and these
patients were divided for analysis into series I (the first 41
patients) and series 2 (the second 31 patients). Reoperations
for MG without thymoma were done in eight patients who had
incapacitating weakness after earlier transcervical or transster-
nal operations (group B). Fifteen patients had MG and
thymoma (group C).
Patient characteristics. The diagnosis of MG was made on
clinical grounds (fluctuating weakness of ocular muscles and
varying combinations of weakness of facial, oropharyngeal,
neck, limb, and respiratory muscles), with unequivocal
response to edrophonium or neostigmine. A decremental
response to repetitive nerve stimulation and abnormal titers of
AChR antibodies were confirmatory," but not required for
diagnosis. There were 27 men and 45 women in group A, two
The Journal of
Thoracic and Cardiovascular
Surgery
and six in group B, and seven and eight in group C. The ages
ranged from 9 to 66 years (68';7, younger than 35) in group A.
from 16 to 54 in group B, and from 25 to 73 years (40';7, older
than 60) in group C. Severity of symptoms, duration of
disease, preoperative AChR titers, and therapy before opera-
tion are recorded in Table I.
Timing of operation and perioperative care. The timing of
operation and selection of drugs before operation were not
uniform. Attempts were made to have all patients free of
aspiration or respiratory symptoms at the time of operation.
but a few who failed to respond to plasmapheresis or other
therapy were operated on while still supported by a respirator
or with the expectation of requiring a tracheostomy early
postoperatively. Immunosuppressive drugs (corticosteroids or
azathioprine) were given preoperatively by the referring
physician in some instances and by us to patients who failed to
respond adequately to anticholinesterase drugs or plasmapher-
esis. Starting in 1982, we gave four to five plasma exchanges
preoperatively to patients who had symptomatic dysphagia.
dysarthria, or respiratory weakness (assessed by standardized
pulmonary function studies'). Postoperatively, the patients
were cared for in the surgical-anesthesia intensive care unit.
Patients who were not extubated in the operating room were
considered for extubation when they met criteria described
previously. '
Classification (severity and results). Preoperatively severi-
ty of symptoms was determined by the patient's maximum
symptoms at any time before operation and classified as
follows: class I (ocular)---ocular signs without generalization:
class II (mild)-mild generalized weakness without oropha-
ryngeal symptoms; class III (moderate)-moderate general-
ized weakness with nonincapacitating oropharyngeal or
respiratory symptoms; class IV (severe)-severe generalized
weakness and/or incapacitating oropharyngeal or respiratory
symptoms; and class V (crisis)-need for intubation and
mechanical ventilation.
Postoperative results were classified as follows: remission-
no signs or symptoms of MG and no medication: asymptom-
atic (A I)-may have minimal ptosis but otherwise clinically
normal, receiving 60 to 240 mg pyridostigmine daily; asymp-
tomatic (A 2)----clinica lly normal when receiving immunosup-
pressive drugs; improved-better than preoperatively when
receiving medication but not A, or A,: unchanged; worse;died
of MG.
Thymic pathologic classification. Hyperplasia was
defined by the presence of lymphoid follicles not ordinarily
seen in the normal adult thymus; involution by the degree of
fatty replacement of thymic tissue. In cases of complete
involution, epithelial thymic remnants were the only indication
that the tissue did represent thymus.'
Method of analysis. The data were analyzed uncorrected
Volume 95
Number 5 "Maximal" thymectomy 749
May 1988

Table I. Patient characteristics

Group A (n = 72) Group B (n = 8) Group C (n = 15)
Maximum severity class (No. of patients)
I. Ocular o o o
II. Mild 13 (18%) o o
III. Moderate 37 (51.5%) I 5
IV. Severe
V. Crisis
I~} (30.5%) ~} (87%) ~} (60'!r)
Mean duration of symptoms 2.8 14.1 1.7
in years (range) (01-20) (1-25) ( 1-8)
Mean preop. AChR titers 73 59 62
(range) (0.6-2000) (4-150) (5-174)
Therapy before operation (No. of patients)
None o o o
Cholinergic agents 65 8 15
Plasmapheresis 30 5 7
Steroids 21 4 9

and corrected for length of follow-up. To assess the effect of Table II. Results of thymectomy (uncorrected)
length of follow-up and the significance of the variables, we
performed multivariate analysis by the Cox proportional Group A Group B Group C
hazard model.' Life table analysis was performed by the (n = 72) tn = 8) in = 15)
Kaplan-Meier method.' As a result of these analyses, some of % No. % No. % No.
the current interpretations differ from an earlier report."
Remission 46 33 12.5 I 13 2
Results Asymptomatic 33 24 25 2 6.5
(AI)
Clinical response. Evaluation was performed at least Asymptomatic 11 8 25 2 46.5 7
1 year after operation. However, of the 72 patients in (AJ
group A, six who were not in remission were lost to Improved 6 4 25 2 20 3
follow-up within the first 11 months. The mean duration Unchanged 4 3 12.5 I 0
Worse 0 0 0
of follow-up was 40 months (6 to 89 months). The
Died (MG) 0 0 13 2
uncorrected results at 6 to 89 months after thymectomy
for groups A, B, and C are recorded in Table II. Life
table analyses for group A are shown in Figs. 1
and 2. older than 60 years and comparable percentages of men
In Group A. the uncorrected results for the 72 (55%) and women (54%) were in remission. By Cox
patients at 6 to 89 months revealed that 96% (69) had regression analysis, a significant difference could not be
benefited from operation; 79% (57) had no symptoms, shown for duration of symptoms before operation,
46% (33) were in remission, and 33% (24) were although 51% of patients (30/59) with symptoms of less
asymptomatic (AI)' Fifty-five percent of those in remis- than 60 months' duration were in remission compared
sion and 67% of those asymptomatic (AI) had achieved with 23% (3/13) with symptoms for 60 months or
that status within 2 years; 60% and 79% by 4 years. Life more.
table analysis for the 72 patients gave a remission rate of The presence or absence of thymic hyperplasia (most
81% at 89 months. Comparison of patients in series 1 in patients less than 36 years of age) or involution did
(N = 41) with those in series 2 (N = 31) indicated that not have a predictive influence. Preoperative AChR
remission rates were significantly higher (p = 0.02) for antibody titers were not related to outcome or thymic
those operated on after June 1982 (series 2), Patients in lesions: Postoperatively, antibody titers fell 25% or more
class II or III had a better response (p = 0.04) than in 35 of 70 patients so studied; 18 of these 35 were in
those in class IV or V. remission or asymptomatic (AI)' The outcome was also
Multivariate analysis failed to show significant differ- excellent in the 17 of the 72 patients whose test results
ences for age, sex, preoperative duration of symptoms, showed seronegativity (eight were in remission and three
thymic hyperplasia and involution, or use of steroids or asymptomatic [AI])' although the number was too small
plasmapheresis preoperatively. Four of five patients to demonstrate statistical significance.

7 50 Jaretzki et al.
100
90
80
70
60
"<f
50
(JJ
c
0 40
'00
(JJ
'E 30
-.--
Ql
a: 20
----
10] •
10 20 30 40
Follow-up (months)
Fig. 1. Remission rates for MG with no thymoma (life table analysis). Remission rates are shown for all patients
in group A and for the subgroups series I and series 2. There was a statistically significant difference between series
1 and series 2 (p = 0.02).
There were two relapses: One was minor, occurred
after 2 years of remission, and lasted I day; the patient
required no therapy and subsequently has been in
remission for 3 years. The other relapse occurred 2 years
after the onset of remission; the patient was treated by
the reinstitution of steroids and 3 years later had become
asymptomatic (A 2) (the patient is not recorded as being
in remission).
In group B, residual thymus was found in all surgical
specimens (2 to 23 gm) regardless of the original
surgical approach, although most computed tomogra-
phic scans were deemed normal or inconclusive. In the
four patients who had previous transcervical procedures
(all were performed elsewhere), thymus was found in
the neck in one of two patients undergoing reexploration
and in the mediastinum in all four; in the four patients
who had previous transsternal procedures (two at
Columbia-Presbyterian Medical Center before 1973),
thymus was found in the neck in all four and in the
mediastinum in three. Reoperation was followed by
considerable improvement in all but one patient, who
remained unchanged (followed up for only I year). Of
the patients in group C, MG with thymona (all stage I),
64% were symptom free; two died 2 and 4 years
postoperatively in crisis (at other institutions).
Pathologic study. Thymus was found in all speci-
mens regardless of the patient's age. In group A, 26
patients had hyperplasia, 28 involution, 17 both, and I
neither. In groups B and C the numbers, respectively,
were 1, 4, 3, and 0 and 3, 6, 5, and 1. The thymoma cell
type was predominantly lymphoid in two patients,
predominantly epithelial in five, with mixed lymphoid
and epithelial elements in eight; no spindle cell thymo-
The Journal of
Thoracic and Cardiovascular
Surgery
..
_ GroupA (n~ 72)
Series # 1 (n=41)
Series #2 (n~31)
50 60 70 80 90
mas were seen. The tumors were mostly solid; some were
cystic and one had a totally calcified shell. All IS were
Bergh classification stage I," without pericapsular inva-
sion, invasion of adjacent organs, or pleural or diaphrag-
matic implants. In one patient, there were two distinct
thymomas of different cell types in separate mediastinal
thymic lobes. Thyroid specimens were submitted from
25 patients; there was thyroiditis in seven (one Hashi-
moto's type), nontoxic nodular goiter in five, and 12
were normal. One specimen showed an area of C-cell
hyperplasia 12; test results after subsequent thyroidecto-
my were negative for thyroid carcinoma.
Complications. There were no operative or hospital
deaths, phrenic or recurrent nerve injuries, or hypopara-
thyroidism in this series or in the 157 patients operated
on since 1973. There were seven major postoperative
complications: bilateral staphylococcal empyema in a
diabetic patient who had been taking corticosteroids and
had presternal folliculitis at the time of operation-the
empyema responded to antibiotic therapy alone; a deep
sternal wound infection in a patient who had a function-
ing tracheostomy at the time of thymectomy and was
receiving corticosteroids-the sternal infection re-
sponded to debridement with myocutaneous closure;
postpericardiotomy syndrome in two patients-both
responded promptly to indomethacin therapy; bilateral
chylothorax in a patient who had had an earlier
transcervical thymectomy-the chylothorax responded
to chest tube drainage; sternal wound dehiscence in a
64-year-old patient with thymoma who was receiving
steroids-after reclosure, with parasternal vertical wire
reinforcement, the sternum healed without complica-
tions; and a pulmonary embolism in a 63-year-old
Volume 95
Number 5
May 1988
100
90
80
70
-:§2. 60
e;
C/l 50
c
a
'00 40
C/l
E
Q)
30
a: 20
10
.-------.-
10 20 30 40
Follow-up (months)
Fig. 2. Remission rates for MG with no thymoma (life table analysis). Remission rates are shown for all patients
in group A and for the subgroups class II or III and class IV or V. There was a statistically significant difference
between the two subgroups (p = 0.04).
patient on postoperative day 8-this was treated with
anticoagulant therapy without further complications.
There were five late deaths. In group A, myocardial
infarction occurred at 3 years in a patient whose MG
had improved, and carcinoma of the lung occurred in a
patient in remission at 5 years. In group B, a patient
whose condition had improved died in a motorcycle
accident 2 years after reoperation. In group C, two
patients died in crisis at other institutions at 2 and 4
years postoperatively without evidence of recurrent
thymoma.
Discussion
With McQuillen,13 we believe the most reliable mea-
sur" of response to thymectomy is complete and sus-
tained remission. The results of patients who are in
worse condition, have recurrent symptoms, require reo-
peration, or die in crisis must also be recorded because
they represent failure of therapy. In this series of 72
patients with MG without thymoma, 30% of whom were
in class IV or V, the uncorrected remission rate at 6 to
89 months was 46%; it would have been 57.5% if six
patients had been excluded (lost to follow-up before I
year and not in remission) and if five patients had been
included (known to have gone into remission after the
study was completed). The remission rate by life table
analysis was 81 % at 7 years. None were in worse
condition, required reoperation, or died in crisis. One
patient had a relapse after a remission lasting 2 years,
subsequently remained asymptomatic on steroids (A 2) ,
and has not been recorded as in remission.
As in other series, age, sex, and cell type of the
thymus (exclusive of thymoma) did not influence out-
"Maximal" thymectomy 75 1
_e - - Group A (n=72)
Class II-Ill (n=50)
Class IV-V (n=22)
50 60 70 80 90
come.":" As we" previously reported, the absolute
AChR antibody titer or even the absence of detectable
antibodies (seronegativity) did not predict outcome.'?
We have operated on patients as young as eight years
(others have done so on patients as young as 2 V2 years")
and as old as 75 years. We believe that age per se is not a
contraindication to operation, operation frequently being
preferable to severe MG or corticosteroids at both
extremes of age. Older patients invariably have residual
thymic epithelium (all in our series and all in an autop-
sy study of accident victims 60 to 90 years of age"),
and although at higher risk they tolerate the maxi-
mum procedure and respond as well as younger indi-
viduals.
Others have found that the milder or shorter the
duration of symptoms and the longer the postoperative
follow-up, the more favorable the results. 14. IS. 18-20 In this
series, the milder cases had a more favorable outcome,
and the longer the follow-up, the better the results.
However, we did not find a statistical difference for
preoperative duration of symptoms, although the uncor-
rected remission rates were much higher for duration of
symptoms less than 60 months. Accordingly, in the
evaluation of the results of different surgical techniques,
allowance must be made for the severity of the symp-
toms, the length of follow-up, and perhaps preoperative
duration of symptoms.
The life table remission rate was better in patients
operated on after June 1982. This finding remains
unexplained. Although the patients in series 2 had
*O'Toole C. Pushparaj N. Studies of aging. 1987. Unpublished
data.

752 Jaretzki et al.
60
D Low
mI Mean
•
50
High
40
~
e;
UJ
c 30
0
'eUJn
'E 20
OJ
a:
10
0
Cervical Sternal
(6) (6)
Type of Thymectomy
Fig. 3. Comparative remission rates for MG with no thymoma (uncorrected). Collated from 18 reports (six
transcervical, six classic transsternaJ. two extended, and four maximal thymectomies: see text for references). The
difference between the maximal and cervical procedures is highly significant (p = 0.0001).
milder symptoms (15% were in class IV or V compared
with 25% in series 1) and although many more received
plasmapheresis preoperatively (84% compared with 29%
in series 1), multivariate analysis failed to show that
these or other parameters analyzed were statistically
significant. However, the combination of operating on
patients with milder symptoms and a more aggressive
exploration of the neck (when the importance of neck
exploration became apparent}' may account, in part,
for these improved results.
The results of maximal thymectomy are better than
those reported for classic transsternal or transcervical
procedures (Fig. 3) and imply that the less residual
thymus, the better the outcome. The reported uncor-
rected remission rates for transcervical thymectomy
have been 6% to 24% (mean 15.7%)21.26 and for the
classic transsternal procedure, 23% to 37% (mean
35%),18.19.22.27.29 whereas those for the maximal proce-
dure are 46% to 58% (mean 51.4%).J5 30. 31 The differ-
ence between the uncorrected remission rates for maxi-
mal thymectomy and transcervical thymectomy is statis-
tically significant (p = 0.0001), if these individual
studies can be compared. Even though, to date, the
uncorrected results of "extended" procedures'<"
approximate those of maximal resection, we prefer the
maximal type of operation because our anatomic study'
indicates that it should leave less thymus in the neck and
the mediastinum, and life table analysis of comparable
patients corrected for length of follow-up should demon-
strate a clear superiority of the maximal procedure.
We believe reoperation, with maximal thymectomy,
The Journal of
Thoracic and Cardiovascular
Surgery
Extended Maximal
(2) (4)
is indicated if a patient with severe MG fails to respond
to one of the more limited resections (or responds and
then deteriorates). The decision is based on clinical
assessment and evaluation of the extent of the original
procedure. In our 15 reoperations (eight in group Band
seven after this series), 10 of the patients had had
previous transcervical and five transsternal thymectomy;
all had residual thymus, even though preoperative
computed tomographic scan had been normal or incon-
clusive in most. Although the mean duration of symp-
toms at reoperation was 14.6 years and all but one
patient had disabling symptoms at the time of operation,
all are markedly improved at this time, including the
patient (Table II) whose condition at I year was
unchanged. The slower response is presumably related
to the severity and duration of the symptoms in this
group compared with those who had a maximal type of
operation initially.
These reoperations show that thymus can be over-
looked both in the neck and mediastinum regardless of
the initial approach unless a maximal type procedure is
used, and the findings imply that as little as 2 gm of
residual thymus can be responsible for severe symptoms,
because subsequent removal has resulted in marked
improvement. Although we did not reexplore the neck in
four of the patients who had previous transcervical
operations, we now believe the reoperative procedure
should also be maximal in extent in the neck and the
mediastinum regardless of which site had been previous-
ly explored.
"The Case Against Transcervical Thymectomy?" no
Volume 95
Number 5
May 1988
longer seems to be a matter of speculation. In our
opinion, now shared by Kirschner* (one of the original
transcervical proponents) and expressed by Henze and
associates" and Fischer and, associates," transcervical
thymectomy should be abandoned in the treatment of
MG without thymoma. In addition, we believe it is
contraindicated in the treatment of thymoma with or
without MG. The remission rates have been significant-
lylower, 27% and 32% of the patients in two series have
required reoperations.F" residual thymus has been
found in all patients, 16.34-37 and there has been subse-
quent clinical improvement in most." Papatestas and
associates," with life table analysis for the transcervical
operations they personally performed, reported remis-
sion rates at 10 years of 40% for all patients and of 64%
for those with mild symptoms and symptoms of short
duration. These figures compare unfavorably with ours.
These authors state that their results were equal to
transsternal thymectomy; however, they compared
transcervical thymectomy with outdated extra pleural
transsternal procedures that were also incomplete
including "median sternotomy (partial or complete),
lateral thoracotomy, or parasternal mediastinotomy.'?"
Also, they have not indicated how many of the patients
who were not in remission (40% to 60% of the total) did
not improve, were worse, or required reoperation.
The transcervical technique recently described by
Cooper.31 which reportedly gives better exposure for
more extensive thymic excision, gave an uncorrected
remission rate of 52% at follow-up (mean 3.5 years) and
suggests he may be removing significantly more thymus
through the neck than is accomplished by others.
However, the anatomy of the gland precludes complete
removal predictably through the neck regardless of the
exposure. Since his patients were not consecutive and
had predominantly mild MG (two of 65 ocular alone,
15% no medication before operation, and twice as many
as we had with mild symptoms), they formed a more
favorable cohort and would thereby be expected to
surpass our results if the resections were equivalent. In
contrast, he reported 44% worsening postoperatively and
one death in crisis.
The occasional mediastinal thymoma that develops
after a transcervical thymectomy," the discovery of
thymomas at transcervical operations" with possible
spreading of the tumor by incomplete resection.v" and
the finding of two distinct thymomas of different cell
"Kirschner PA. Discussion of Papatestas AE et al. Effects of
thymectomy in myasthenia gravis. New York Surgical Society
meeting, Jan. 8, 1986. Unpublished data.
"Maximal" thymectomy 75 3
type in separate mediastinal lobes* are additional rea-
sons for abandoning the transcervical approach. The
morbidity and mortality of the transsternal procedures
are no longer deterrents and the cosmetic benefit of the
cervical incision is offset by the lower rates of remission
or asymptomatic (AI) status after partial thymectomy
and by the patients who continue to have disabling
symptoms. Although some believe that the transcervical
procedure should be done first and that extensive
procedures should be reserved for patients whose condi-
tions fail to improve after the transcervical procedure,
that policy would condemn many patients to persistent
symptoms, some severe.
The incidence of thymoma in this series was 16% and
all cases were nonmalignant (stage I). In our overall
experience of 154 consecutive thymectomies, the inci-
dence of thymoma was 17%. Of the 154 patients, 2.3%
had invasive thymoma and the likelihood of a thymoma
being present increased with age (3% for 20 years and
younger; 12% for 21 to 45 years; and 35% for 46 and
older t). Although many patients with thymoma respond
symptomatically to thymectomy, and thymectomy is
indicated, it has long been recognized that the response
is less favorable than in patients without thymo-
ma.20.41.43.44 In our patients with thymoma, the outcome
was less favorable than in those without thymoma, and
the delayed mortality from MG was 13% (2/15)
without tumor recurrence. Even so, almost two thirds
did become symptom free. Although we use postopera-
tive radiation for stage II and III thymoma, we have not
used it for stage I, contrary to the recommendations of
Mondon and colleagues."
Complete removal of coexisting thymus gland (maxi-
mal thymectomy) is also indicated in the presence of
thymoma with or without MG. This recommendation is
supported by the presence of hyperplasia with variable
degrees of involution in the thymus in association with
the thymoma in all but one in this series and the
occasional late onset of MG (6% to 10%) after removal
of an asymptomatic thymoma without removal of the
coexisting thymus."
We are indebted to Robert R. Sciacca, EngScD, Depart-
ment of Medicine, College of Physicians and Surgeons,
Columbia University, New York, for the statistical analysis.
"Wolff M, Lovelace RE. Jaretzki A III. Two histologically dissimilar
thymomas in a patient with myasthenia gravis. Unpublished
data.
tEllis K, Austin JHM, Jaretzki A III. Radiologic detection of
thymoma in patients with myasthenia gravis. Unpublished data.

754 Jaretzki et al.
REFERENCES
I. Wekerle H, Muller-Herrnelink HK. The thymus in
myasthenia gravis. Curr Top Pathol 1986;75: 179-206.
2. Engel AG. Myasthenia gravis and myasthenic syndrome.
Ann Neurol 1984;16:519-34.
3. Jaretzki A III, Bethea M, Wolff M, Olarte MR, Penn
AS, Rowland LP. A rational approach to total thymecto-
my in the treatment of myasthenia gravis. Ann Thorac
Surg 1977;24: 120-30.
4. Masaoka A, Nagaoka Y, Kotake Y. Distribution of
thymic tissue in the anterior mediastinum: current proce-
dures in thymectomy. J THoRAc CARDIOVASC SURG 1975;
70:747-54.
5. Jaretzki A III, Wolff M. "Maximal" thymectomy for
myasthenia gravis: surgical anatomy and operative tech-
nique. J THoRAc CARDIOVASC SURG (In press).
6. Penn AS, Schotland DL, Lamme S. Antimuscle and
antiacetylcholine receptor antibodies in myasthenia gra-
vis. Muscle Nerve 1986;9:407-15.
7. Younger DS, Braun NMT, Jaretzki A III, Penn AS,
Lovelace RE. Myasthenia gravis: determinants for inde-
pendent ventilation after transsternal thymectomy. Neu-
rology 1984;34:336-40.
8. Penn AS, Jaretzki A III, Wolff M, Chang HW, Tenny-
son V. Thymic abnormalities: antigen or antibody?
Response to thymectomy in myasthenia gravis. Ann NY
Acad Sci 1981;377:789-91.
9. Lawless JF. Statistical models and methods for life table
data. New York: John Wiley & Sons, Inc., 1982.
10. Younger DS, Jaretzki A III, Penn AS, et a!. Maximum
thymectomy for myasthenia gravis. Ann NY Acad Sci
1987;505:832-5.
11. Bergh NP, Gatzinsky P, Larrsons S, et a!. Tumors of the
thymus and thymus region. I. Clinicopathological studies
on thymomas. Ann Thorac Surg 1978;25:91-8.
12. LiVolsi VA. Calcitonin: the hormone and its significance.
In: Fenoglio CM, Wolff M, eds. Progress in Surgical
Pathology; vol 1. New York: Masson Pub!. USA Inc.,
1980:88-9.
13. McQuillen MP. Symposium on therapeutic controversies:
myasthenia gravis-thymectomy. Trans Am Neurol
Assoc 1978; I03:283-6.
14. Mulder DG, Hermann C Jr, Keesey J, Edwards H.
Thymectomy for myasthenia gravis. Am J Surg 1983;
146:61-6.
15. Olanow CW, Wechsler AS, Roses AD. A prospective
study of thymectomy and serum acetylcholine receptor
antibodies in myasthenia gravis. Ann Surg 1982; 196:113-
21.
16. Stump WJ, Adornato BT, Engel WK, McIntosh CL,
Castleman BJ. Thymectomy in myasthenia gravis. Neu-
rology 1978;28:372-3.
17. Soliven BC, Lange DJ, Penn AS, Younger D, Jaretzki A
III, Rowland LP. Seronegative myasthenia gravis. Neu-
rology (In press).
The Journal of
Thoracic and Cardiovascular
Surgery
18. Mulder DG, Herrmann C, Buckberg GD. Effect of
thymectomy in patients with myasthenia gravis: a sixteen
year experience. Am J Surg 1974; 128:202-5.
19. Emeryk B, Strugalska MH. Evaluation of results of
thymectomy in myasthenia gravis. J Neurol 1976;
21 I :I 55-68.
20. Papatestas AE, Genkins G, Kornfeld P, et al. Effects of
thymectomy in myasthenia gravis. Ann Surg 1987:
206:79-88.
21. Masaoka A, Monden Y. Comparison of the results of
transsternal simple, transcervical simple and extended
thymectomy. Ann NY Acad Sci 1981;377:755-65.
22. Matell G, Lebram G, Osterman PO, Pirskanen R. Follow
up comparison of suprasternal vs transsternal method for
thymectomy in myasthenia gravis. Ann NY Acad Sci
1981 ;377:844-5.
23. Donnelly RJ, Laquaglia MP, Fabri B, Hayward M,
Florence AM. Cervical thymectomy in the treatment of
myasthenia gravis. Ann R Coli Surg Engl 1984;66:
305-8.
24. Paletto AE Maggi G. Thymectomy in the treatment of
myasthenia gravis: results in 320 patients. Int Surg
1982;67:13-6.
25. Papatestas AE, Genkins G, Kornfeld P. Comparison of
the results of transcervical and transsternal thymectomy
in myasthenia gravis. Ann NY Acad Sci 1981;377:766-
78.
26. Klingen G, Johansson L, Westerholm CJ, Sundstroom C.
Transcervical thymectomy with the aid of mediastinosco-
py for myasthenia gravis: eight years' experience. Ann
Thorac Surg 1977;23:342-7.
27. Buckingham JM, Howard FM Jr, Bernatz PE, et al.
The value of thymectomy in myasthenia gravis: a
computer-assisted matched study. Ann Surg 1976; 184:
453-8.
28. Masaoka A, Maeda M, Monden Y, Nakahara K, Kotake
Y. Clinical results following thymectomy in myasthenia
gravis. Med J Osaka Univ 1976;26: 117-28.
29. Oosterhuis HJ. Observations of the natural history of
myasthenia gravis and the effect of thymectomy. Ann
NY Acad Sci 1981;377:678-90.
30. Rubin JW, Ellison RG, Moore HV, Pai GP. Factors
affecting response to thymectomy for myasthenia gravis. J
THORAC CARDIOVASC SURG 1981;82:720-8.
31. Fischer JE, Grinvalski HT, Nussbaum MS, et a!. Aggres-
sive surgical approach for drug-free remission from myas-
thenia gravis. Ann Surg 1987;205:496-503.
32. Ferguson TB, Hankins JR, Mayer RF, Satterfield JR.
Thymectomy for myasthenia gravis: 14 year experience.
Am Surg 1985;201:618-25.
33. Jaretzki A III. Myasthenia gravis: the case against
transcervical thymectomy. Trans Am Neurol Assoc
1978;103:288-91.
34. Henze A, Biberfeld P, Christensson B, Matell G, Pirskan-
en R. Failing transcervical thymectomy in myasthenia
Volume 95
Number 5
May 1988
gravis: an evaluation of transsternal re-exploration. Scand
J Thorac Cardiovasc Surg 1984; 18:235-8.
35. Masaoka A, Monden Y, Seike Y, Tanioka T, Kagotani
K. Reoperation after transcervical thymectomy for myas-
thenia gravis. Neurology 1982;32:83-5.
36. Rosenberg M, Jauregui WO, DeVega ME, Herrera MR,
Roncoroni AJ. Recurrence of thymic hyperplasia after
thymectomy in myasthenia gravis: its importance as a
cause of failure of surgical treatment. Am J Med
1983;74:78-82.
37. Rosenberg M, Jauregui WO, Herrera MR, Roncoroni
AJ, Rojas OR, Olmedo GSM. Recurrence of thymic
hyperplasia after transsternal thymectomy in myasthenia
gravis. Chest 1986;89:888-9.
38. Pirskanen R, Matell G, Henze A. Results following
transsternal thymectomy after failing transcervical "thy-
mectomy." Ann NY Acad Sci (In press).
39. Cooper J. An improved technique for transcervical thy-
mectomy in myasthenia gravis. Ann Thorac Surg (In
press).
40. Austin EH, Olanow CW, Wechsler AS. Thymoma
following transcervical thymectomy for myasthenia gra-
vis. Ann Thorac Surg 1983;35:548-50.
41. Shamji F, Pearson FG, Todd TRJ, Ginsberg RJ, Ives R,
Cooper JD. Results of surgical treatment of thymoma. J
THORAC CARDIOVASC SURG 1984;87:43-7.
42. Kirscher PA. Discussion of Shamji and associates:"
43. Simpson JA. An evaluation of thymectomy in myasthenia
gravis. Brain 1958;81: 112-45.
44. Perlo VP, Poskanzer DC, Schwab RS, Viets HR, Osser-
man KE, Genkins G. Myasthenia gravis: evaluation of
treatment in 1355 patients. Neurol 1966;16:431-9.
45. Mondon Y, Nakahara K, Iioka S, et al. Recurrence of
thymoma; clinicopathological features, therapy and prog-
nosis. Ann Thorac Surg 1985;39:165-9.
46. Namba T, Brunner NG, Grob D. Myasthenia gravis in
patients with thymoma, with particular referrence to
onset after thymectomy. Medicine 1978;57:411-33.
Discussion
Dr. Donald G. Mulder (Los Angeles. Calif). Dr. Jaretzki
has made a point I think we would support completely, that a
median sternotomy is the appropriate approach for removing
the thymus.
I would have some reservations, however, about seeing this
extended procedure, the maximal thymectomy, recommended
as the accepted operation in these patients. I say that for two
reasons. First, the increased mobidity associated with dis-
secting high in the anterior and lateral mediastinum and neck
with regard to phrenic nerve and even recurrent nerve injury is,
I think, substantial. This is a catastrophic injury in a sick
myasthetic patient who often already has respiratory compro-
mise. Second, I am not sure the good results presented are due
to the extension of the more standard operation. I would like to
review briefly our previously reported experience' over the past
27 years with such patients (Am J Surg 1983;146:61-6),
"Maximal" thymectomy 755
because I think our results are comparable to those presented
today.
There were a total of 781 patients in this group of which 249
underwent thymectomy. Approximately 20% had thymoma,
which is similar to the experience of Dr. Jaretzki.
The patients were categorized according to the severity of
disease and only a small number, nine, were in the ocular stage
or class I, whereas the majority were in the same category as
the patients of Dr. Jaretzki, namely classes II and III, which
are the more severely ill.
Thymectomy was done through a median sternotomy in
every instance, and the entire thymus including the cervical
pedicles was removed. The pleural spaces were entered when-
ever the gland extended laterally to ensure careful dissection
near the phrenic nerves.
The patients were followed up continually, often yearly,
with an average follow-up of 7'/2 years. The remission rate
(defined as no medication and no symptoms) was 51%.
Improvement (less medication with better control of the
symptoms) was found in an additional 36%, so that the overall
benefit from operation was 87%.
The 51 patients with thymoma did not fare quite as well.
Even so, the remission rate was 37% with an improvement
noted in 31% for an overall benefit of 68%.
By contrast, in the 198 patients without thymoma, a
remission rate of 54% was noted along with an improvement of
37% for an overall benefit from operation of 91%.
We do believe that the severity of illness preoperatively
made a difference in the result, because the rate of remission
dropped from 67% in class I to 25% in class IV. These were the
patients with longstanding myasthenia who often had invasive
tumor, and they did not do well.
On the basis of this experience, then, I have reservations
about recommending this extended procedure with its
increased risk as being either necessary or even advisable.
However, I would like to thank Dr. Jaretzki and his group for
continuing to look for better solutions to this difficult prob-
lem.
Dr. Yasunaru Kawashima (Osaka. Japan). I would like to
congratulate Dr. Jaretzki and his co-workers for the excellent
results obtained by the maximal thymectomy procedure. I also
definitely recommend this kind of thymectomy in the treat-
ment of MG.
We have performed thymectomy in 297 patients with MG
at Osaka University Hospital. In the early days, we also
performed transsternal simple thymectomy and transcervical
thymectomy. However, we have done the extended thymecto-
my procedure exclusively since Masaoka, one of my co-
workers, found histologically in more than 70% of patients that
extracapsular thymic tissue existed among the fat surrounding
the thymus.
Since April 1973, 173 of our patients without thymoma and
63 patients with thymoma underwent extended thymectomy.
In this procedure, the adipose tissue in the vicinity of the
thymus is resected as much as possible until the major blood
vessels of the anterior mediastinum, pericardium, pleural
surface, and phrenic nerve are almost completely exposed.
The remission rate of the patients undergoing extended
thymectomy increased with passage of time after operation,
but there were no statistically significant differences between
the remission rates of extended thymectomy and of other

7 5 6 Jaretzki et al.
methods of thymectomy. These rates were 40% to 45% 10
years after operation.
However, the palliation rate, which I presume to be
equivalent to the asymptomatic and remission rates reported
by Dr. Jaretzki, was much higher, particularly in patients who
underwent extended thymectomy. The palliation rate
increased with passage of time after operation regardless of the
method of thymectomy. There were statistically significant
differences between palliation rates in patients who underwent
extended thymectomy and in patients who underwent simple
or cervical thymectomy.
There was a limited difference in the palliation rates of
extended thymectomy between the patients with and without
thymoma. However, the remission rate in patients with
thymoma remained low for a long time after operation and
was statistically significantly different from that in patients
without thymoma.
We have been recommending extended thymectomy for the
treatment of MG for many years and so far have been satisfied
with the result. We can obtain an excellent surgical view
through a median sternotomy up to the thyroid gland.
I would like to ask Dr. Jaretzki whether he has found any
thymic tissue in cervical fat that was not accessible through the
median sternotomy. I am wondering whether or not we should
increase our extended thymectomy to maximal thymectomy.
Dr. Joel D. Cooper (Toronto. Ontario. Canada). I too
support the concept of maximal thymectomy, but would
disagree with Dr. Jaretzki on the type of surgical approach
required to produce it.
We do a form of transcervical thymectomy. It uses a small
incision, but it is a big operation. It involves removing the
thymus gland, pericardium, and pleura when necessary, and
the pleural fat-all under direct vision through a small neck
incision. We have a series of 53 consecutive patients with
longest follow-up only to 5 years; of course with time, as
everyone has pointed out, the results would improve. These
patients had a rather severe form of illness beforehand:
Forty-four percent were receiving prednisone and pyridostig-
mine (Mestinon). The mean follow-up grade has been 0.5
(zero is asymptomatic, I is ocular symptoms only, 2 is mild
generalized weakness, and 3 is moderate generalized weak-
ness). There was improvement of one or more grades in 95% of
the patients and of two or more grades in 81%; 6% showed no
improvement. The complete remission rate, given the duration
of follow-up, was equivalent to what Dr. Jaretzki has shown on
his time curve. Dr. Jaretzki's anatomic drawings have been
important in permitting a complete thymectomy to be (done
with) this particular approach.
I would also point out that if complete remission were an
indication of complete thymectomy, we would have a right to
expect a complete remission in every patient in which Dr.
Jaretzki's operation is done., but unfortunately the nature of
the disease is different from that. I think we must rather be
able to compare results in patients with similar stages of
disease and followed up for a similar period of time.
Finally, thymectomy through the transcervical incision is a
specialty that you have to work at if you want to accomplish it.
I really only wanted to make the point that equivalent results
can be obtained by performing complete thymectomy through
a variety of different approaches. Although I do not currently
practice in the United States, I have heard about something
here called diagnosis related groups (DRGs), and I frankly
The Journal of
Thoracic and Cardiovascular
Surgery
would have thought that the opportunity of achieving equiva-
lent results with a procedure that allows the patient to go home
the next day might have some attraction at centers in the
United States.
Dr. Jaretzki (Closing). I appreciate the discussers' com-
ments. Dr. Mulder's warning concerning the risks of the
thymic dissection that we have described is appropriate; the
maximal operation as we perform it is time-consuming and has
potential for harm. There must be no injury to the phrenic or
recurrent nerves; this could be catastrophic, especially to a
patient with MG. The wide mediastinal exposure we obtain by
opening both pleural spaces not only ensures complete thymus
removal but helps safeguard the phrenic nerves. The same
applies to the neck. We have demonstrated that this operation
can be done with safety comparable with that of the standard
transsternal procedure; the morbidity has been low, there has
been no mortality, there have been no phrenic or recurrent
nerve injuries, and the results justify the increased operative
time. If the thoracic surgeon has had only limited experience
doing this type of neck operation, he should seek assistance for
this part of the procedure, initially at least.
It is true that Dr. Mulder's overall uncorrected remission
rates appear to equal those of the maximal procedure, even
though his neck and mediastinal dissections are less complete.
However, the two series are not comparable, and life table
analysis should confirm this. His patients should do better
since milder symptoms and longer follow-up favor a higher
remission rate; 59% of his patients had mild MG (compared to
our 20%) and his mean follow-up time is 7.5 years (compared
to our 3.4 years). Of note, since our data were tabulated. our
uncorrected remission rate has risen to 54% and we anticipate
a continued rise with time.
I appreciate Dr. Kawashima's comments as well. We are
indebted to his group for the demonstration of microscopic foci
of thymus in mediastinal fat outside the capsule of the gland.
We believe that a combination of their observations reported in
1975 (J THoRAc CARDIOVASC SLRG 1975;70:747-54) and ours
reported in 1977 (Ann Thorac Surg 1977;24:120-30) have led
to a better understanding of the anatomy of the thymus and
have led to more extensive resections with improved results. It
is not clear to me, however. why in their hands the extended
procedure has not yielded improved remission rates, because in
all other series we have analyzed there appears to be a direct
relationship between the remission rate and the extent of
resection. In answer to Dr. Kawashima's question regarding
the presence of thymus in the neck, in 40% of our dissections
some thymic tissue was found outside the confines of the
classic cervical lobes.
Although I respect Dr. Cooper's surgical skill and am
indebted to him for his suggestion that I describe our
mediastinal dissection with the term "anterior mediastinal
exenteration," I am in total disagreement with his conclusions.
First, the anatomy of the thymus in the mediastinum (espe-
cially the presence of significant amounts of thymus in the
aortopulmonary window, under the phrenic nerves, and in the
distal mediastinal fat) precludes complete removal through the
neck in the majority of patients, no matter how skilled the
surgeon. Second, the need for as complete removal of thymus
as possible is supported by the following: (1) the prevention of
experimental autoimmune MG in rabbits after neonatal
complete thymectomy, whereas it is not prevented by partial
thymectomy; (2) the demonstration that the thymus has both
Volume 95
Number 5
May 1988
antibody- and antigen-bearing cells; (3) the patients who did
not benefit from incomplete thymectomy responded to reoper-
ation with more complete resections; (4) the demonstration
that as little as 2 to 3 gm of residual thymus can be severely
symptomatic and its removal therapeutic; and (5) the superior
remission rates of comparable patients with comparable
follow-up evaluation after maximal thymectomy as compared
with the transcervical or other less extensive procedures.
Finally, although Dr. Cooper reports a 50% uncorrected
remission rate, he has a more favorable patient cohort and he
also has a 44% incidence of progression of symptoms after
thymectomy. In our series, in addition to the considerably
higher remission rates, we have virtually eliminated the
devastating and life-threatening symptoms that persist indefi-
nitely in many patients with lesser resections.
"Maximal" thymectomy 757
In reference to Dr. Cooper's comments regarding DRGs in
MG, I believe two comments are appropriate. First, we in the
medical profession must take care not to compromise our care
whatever the pressures. Second, the higher remission rates and
marked decrease in the protracted illness in our patients should
reduce the cost of their medical care and greatly offset the few
extra days in the hospital necessitated by a thoracic as
compared with a cervical procedure.
Clearly, much remains unexplained. We can hope that
successful treatment of MG will eventually not require an
operation. However, at this time, analysis of the available data
supports the thesis that the more comprehensive the thymic
resection, the better the results and that a comprehensive
resection cannot be achieved through the neck alone or even
via a limited transsternal procedure.