The Risk of Esophageal Cancer

in Patients With Achalasia

A Population-Based Study
Robert S. Sandler, MD, MPH; Olof Nyr\l=e'\n,MD, PhD; Anders Ekbom, MD, PhD; Glenn M. Eisen, MD, MPH;
Jonathan Yuen, PhD; Staffan Josefsson

Objective.\p=m-\Todetermine more precise and accurate cancer risk estimates for have been made without accurate data
achalasia that could be used to plan surveillance. on the magnitude of cancer risk in acha
lasia. Moreover, while it is theoretically
Design.\p=m-\Cohort.
Setting.\p=m-\Swedishpopulation. possible to detect early cancer using en-
Participants.\p=m-\Allpatients with achalasia listed in the population-based Swed- doscopy, there are only isolated reports
of the success of such a program.8 Esoph
ish Inpatient Register from 1964 through 1989.
Main Outcome Measures.\p=m-\Theobserved number of cancers in the cohort was
ageal cancer is uncommon in the general
population. Even if the relative risk of
compared with expected numbers of cancers (standardized incidence ratio [SIR]) cancer in patients with achalasia is con
for each 5-year age group and calendar year of observation, calculated using data siderable, the absolute risk is still small.
from the Swedish Cancer Registry. A large number of examinations might
Results.\p=m-\Atotal of 1062 patients with achalasia accumulated 9864 years of be required to detect a single cancer.
follow-up. The mean age at entry was 57.2 years, and the mean age at cancer di- We therefore conducted a large popu
lation-based cohort study of patients
agnosis was 71.0 years. Esophageal cancer occurred in 24 patients. The risk of with achalasia that was designed to de
esophageal cancer in the first year after achalasia diagnosis was extremely high termine more precise estimates of can
(SIR, 126.3; 95% confidence interval [Cl], 63.0 to 226.1) as a consequence of cer risk. The information derived from
prevalent cancers leading to distal esophageal obstruction simulating achalasia. such a study could be used to help de
During years 1 to 24, the risk was increased more than 16-fold (SIR, 16.6; 95% CI, sign logical, data-based recommenda
8.8 to 28.3). Annual surveillance after the first year would require 406 endoscopic tions for surveillance.
examinations in men and 2220 in women to detect one cancer.
Conclusions.\p=m-\Patientswith achalasia are at markedly increased risk of devel- METHODS
oping esophageal cancer. A substantial number of surveillance examinations might The data for the current study were
be required to screen for cancers, especially in women. It is not known whether derived from the Swedish Inpatient Reg
surveillance will result in improved survival. ister. Because there is almost no private
(JAMA. 1995;274:1359-1362) inpatient treatment in Sweden, hospital-
provided medical services are, in effect,
population-based and referable to the
ACHALASIA is an esophageal motor have associated esophageal cancer with county in which the patient lives. Be
disorder characterized by aperistalsis long-standing achalasia.1 Whether pa ginning in 1964 and 1965, the National
and failure of the lower esophageal tients with achalasia are at greater risk Board of Health and Welfare started
sphincter to relax on swallowing. As a for esophageal cancer than members of collecting data on individual hospital dis
consequence, the esophagus may become the general population is not known with charges in this inpatient register. Be
dilated and filled with food debris and certainty. Most of the evidence for an sides the national registration number
fluid. For more than 100 years clinicians association between achalasia and can (a unique personal identifier assigned to
cer comes from case reports and small all Swedish residents), each record con
series.2"7 Hospital-based cohort studies tains data on hospital department, pro
that suggest an elevated risk have gen cedures for surgery and anesthesia, and
From the Department of Medicine and Center for
Gastrointestinal Biology and Disease, University of erally been limited by small sample size up to eight discharge diagnoses, coded
North Carolina at Chapel Hill (Drs Sandier and Eisen); or short follow-up.89 Not all the cohort according to the seventh revision of the
Department of Epidemiology, Uppsala (Sweden) Uni- studies have demonstrated an increased International Classification of Diseases
versity (Drs Nyr\l=e'\n,Ekbom, and Yuen and Mr Josefs- risk of cancer.10,11 (ICD-7) through 1968 and the eighth
son); and Department of Epidemiology, Harvard
School of Public Health, Boston, Mass (Dr Ekbom). Some authors and organizations have revision (ICD-8) thereafter. The num
Reprint requests to Division of Digestive Diseases recommended endoscopie surveillance ber of hospitals delivering data to the
and Nutrition, CB 7080, 423A Burnett-Womack Bldg,
University of North Carolina at Chapel Hill, Chapel Hill,
for patients with achalasia to detect early register has increased steadily: in 1969,
NC 27599-7080 (Dr Sandler). cancer.3,12"14 These recommendations the percentage was 75%, reaching 85%

Downloaded From: http://jama.jamanetwork.com/ by a University of Pennsylvania User on 11/10/2015

 by the end of 1983. Since the country is
divided into mutually exclusive hospital
catchment areas and the people obliged
to use the hospital in their catchment
areas, the coverage of the register at
each point in time corresponds to a de
fined population. More detailed descrip
tions of this inpatient register have been
published previously.15,16
The Cohort
All patients recorded in the inpatient
register with a discharge diagnosis of
achalasia (ICD-7 code, 539.01; ICD-8
code, 530.01) were included. The national
registration number was used to select
the first recorded discharge with this
diagnosis for each individual. Patients
who were not admitted to the hospital
are not identified using this approach,
but during the study period, Swedish
patients with achalasia were exclusively
treated by otolaryngologists who used
rigid esophagoscopy, which was always
done on an inpatient basis. As a result,
all new cases of symptomatic achalasia
and all prevalent cases requiring treat
ment were included in the study. Man
datory health insurance with uniformly
low patient cost, excellent availability
of hospital care, and prevailing medical
practice make it likely that virtually all
patients with symptomatic achalasia
were admitted and treated. Patients who
develop esophageal cancer within 2 years
of first discharge with an achalasia di
agnosis are excluded to avoid selection
bias.
Follow-up
Information on deaths through 1989
was obtained by linking nine-digit na
tional registration numbers17 to the na
tionwide Registry of Causes of Death.
The registry includes all deceased per
sons, whether they died in Sweden or
abroad.18 The underlying cause of death
is generally determined from data on
medical death certificates, assigned in
accordance with the internationally es
tablished form. A corresponding link
age to the Register of Population Move
ments identified dates of emigration and
reimmigration, if applicable, among the
cohort members. The national Swedish
Cancer Registry was used to ascertain
all incident cancers diagnosed in the co
hort from the start of follow-up until the
end of 1989. The registry coded malig
nant diseases according to the ICD-7
classification during the entire study pe
riod. Physicians and pathologists or cy-
topathologists submit mandatory re
ports to the cancer registry,19 which was
approximately 98% complete at the clos
ing date of the study.
To remove entries in the cohort with
plausible but erroneous national regis-
Downloaded From: http://jama.jamanetwork.com/ by a University of Pennsylvania User on 11/10/2015
Table 1. Characteristics of the Achalasia Cohort
Men
No. of
Cases
Years of
Observation
No. of
Cases
397
20-39 120 1410 80
60-69 105 842 94
Overall 561 4970 501
tration numbers, thus not correspond
ing to any identifiable person, we also
linked the cohort to the Register of the
Total Population. If a national registra
tion number could not be found in any of
the registers (ie, the Cancer Registry,
Registry of Causes of Death, Register
of Population Movements, or Register
of the Total Population), it was likely
that the number did not correspond to
an existing person, and it was removed.
The time of observation was calcu
lated from the date of entry into the
cohort (date of first discharge with the
achalasia diagnosis) until emigration
(with the addition of person-years after
reimmigration, if any), death, or the end
of the observation period (December 31,
1989), whichever occurred first. The
analysis included second cancers and can
cers found on autopsy, both among acha
lasia cases and in the underlying refer
ence population.
Statistical Methods
We calculated the expected number
of cancers by multiplying the number of
person-years of observation for each sex
by age-specific cancer incidence rates
for each 5-year age group and calendar
year of observation. The expected rates
were derived from the entire Swedish
population using data from the national
Swedish Cancer Registry. The standard
ized incidence ratio (SIR) was used as a
measure of relative risk defined as the
ratio of the observed to expected can
cers. A 95% confidence interval (CI) was
calculated assuming that the observed
cases followed a Poisson distribution.
To assess latency, we also conducted
sex-specific analyses, using 0 to 1 year,
1 to 4 years, 5 to 9 years, and 10 to 24
years of follow-up since diagnosis.
RESULTS
The cohort was composed of 1062 pa
tients with achalasia who accumulated
9864 years of follow-up (Table 1). There
were slightly more men than women in
this cohort (561 men vs 501 women).
The average age at entry was 55.7 years
for men and 59.0 years for women.
Esophageal cancer occurred in 24 pa
tients. The mean age at esophageal can-
Women Overall
Years of No. of
Cases
Years of
Observation
Observation
1046 200 2456
199 1730
4894 1062 9864
cer diagnosis was 71.0 years. Women
were older than men at the time of can
cer diagnosis (mean age, 75.7 vs 70.0
years). The interval between achalasia
diagnosis and cancer incidence was simi
lar in men and women.
Table 2 provides information on the
number of esophageal cancers that de
veloped among members of the cohort
by sex and latency interval. The risk of
esophageal cancer in the first year after
achalasia diagnosis was extremely high
(SIR, 126.3). During the first year, the
risk for esophageal cancer was more than
twice as high in men as women. This
high rate in the first year may be in
flated due to cancers leading to or mis
interpreted as achalasia (so-called sec
ondary achalasia20), although some of
these cancers may represent primary
achalasia that had developed into can
cer. The data are presented simply to
highlight the importance of excluding
prevalent cancers. For longer latency
intervals, the overall SIRs were stable
between 16 and 17. There was no con
sistent trend during the calendar period
(data not shown). The cumulative inci
dence of esophageal cancer, excluding
the first year of follow-up, was 145.9 per
100 000 population. There were 14 squa-
mous cell carcinomas, six adenocarcino-
mas, and four undifferentiated cancers.
If achalasia predisposes to esophage
al cancer, nonesophageal cancers would
be expected to be no more common in
patients with achalasia than in mem
bers of the general population (specific
ity). We therefore calculated the SIR
for other cancers. During the first year
of follow-up, cancers of the pancreas
(SIR, 10.5; 95% CI, 1.2 to 37.9) and stom
ach (SIR, 25.6; 95% CI, 12.2 to 47.0),
particularly cancer of the gastric cardia
(SIR, 159.7; 95% CI, 43.0 to 408.8), were
more common in men. Cancers of the
stomach (SIR, 27.4; 95% CI, 10.0 to 59.4),
particularly cancer of the gastric cardia
(SIR, 515.4; 95% CI, 138.7 to 1319.5),
and cancer of the uterine corpus (SIR,
11.3; 95% CI, 1.3 to 40.5) were more
common in women (P<.05). It is likely
that the achalasia was secondary to some
of these cancers since they were not
more common after the first year of fol-
 Table 2. Observed Cases of Esophageal Cancer, Person-Years of Follow-up, and Standardized Incidence Ratio (SIR) in 1062 Patients With Achalasia by Sex
and Latency Interval
Men Women Overall

Latency Cancer Cancer Cancer

Interval, y Cases Person-Years* SIR (95% Cl)t Cases Person-Years* SIR (95% Cl)t Cases Person-Years* SIR (95% Cl)t
497 146.9(67.0-278.9) 453 77.5 (8.7-280.0) 950 126.3(63.0-226.1)
1357 18.4(3.7-53.9) 1280 14.3(0.2-79.4) 2637 17.2(4.6-44.0)
5-9 23.1 (7.4-53.9) 0 (0-39.6) 16.2(5.2-37.8)
1331 17.7(3.6-51.8) 1441 13.5(0.2-74.9) 2772 16.4(4.4-42.1)
1-24 20.1 (10.0-35.9) 8.4 (0.95-30.5) 8909 16.6(8.8-28.3)
*Leap days were not counted in the person-years calculations.
fCI indicates confidence interval.

low-up. No other cancers were consis
tently more common in the achalasia
cohort. Of special interest are cancers
that could serve as indicators of tobacco
smoking and alcohol consumption. There
was no increased risk of lung cancer (4.0
observed vs 6.1 expected; SIR, 0.7; 95%
CI, 0.2 to 1.7). There was an increased
risk of primary liver cancer that was not
statistically significant (7.0 observed vs
3.0 expected; SIR, 2.3; 95% CI, 0.9 to
4.7). The same was true for pancreatic
cancer (6.0 observed vs 3.1 expected;
SIR, 1.9; 95% CI, 0.7 to 4.2). The overall
SIR (excluding esophageal cancer and
excluding the first year) was 0.96 (95%
CI, 0.8 to 1.2).
Knowledge of cancer risk is impor
tant for prognosis but also might be used
to evaluate potential benefits of surveil
lance. After excluding the first year, we
calculated the number of yearly endo
scopie examinations required to detect
one cancer by dividing the person-years
of observation by the observed cancers
in patients with achalasia. Overall, 681
examinations would be necessary to de
tect one cancer. This number includes
cancers that develop in association with
achalasia and those that develop spon
taneously. The number of examinations
varies considerably by sex. Because the
risk of esophageal cancer is much higher
in men, fewer examinations are needed
in men (406 examinations) than in wom
en (2220 examinations).
COMMENT
This large, population-based cohort
study has demonstrated that patients
with achalasia are at substantially
greater risk of developing esophageal
cancer. The risks are extremely high
during the first year after diagnosis due
to inclusion of prevalent cases. After
the first year, the risk stabilizes at ap
proximately 16-fold. Despite the high
risks, there is currently no evidence that
surveillance of these patients will be
valuable. These data suggest that a large
number of surveillance examinations
would be necessary to detect one can
cer, particularly in women.
The first report of cancer in a patient
with achalasia is attributed to Fagge1 in
1872. Since that time there have been
numerous case reports and reviews.
Most authors attribute esophageal can
cer to stasis, chronic esophagitis, stag
nation of food, and chronic inflamma
tion.21 The diagnosis of esophageal cancer
is often delayed because the symptoms
of cancerdysphagia and weight loss
are easily attributed to the achalasia.13
In our study, the mean survival after
diagnosis was only 0.7 year (range, 0 to
5.2 years), indicating that early diagno
sis was not a common feature. The can
cers are reported to occur at an earlier
age than esophageal cancer in the gen
eral population in some reports22 but not
others.23 The cancers typically have squa-
mous histology and tend to occur in the
middle third.24 The percentage of pa
tients with achalasia who have devel
oped cancer has been reported to be as
high as 29% in hospital-based series.13
As additional case reports were added
to the literature, the weight of their
numbers led to the conclusion that acha
lasia is a risk factor for subsequent can
cer. Unfortunately, it is impossible to
draw firm conclusions from case reports.
They may simply report instances in
which both diseases occurred in the same
patient by coincidence. To properly as
sess risk, it is necessary to compare in
cidence in those affected with achalasia
to the incidence in an unaffected popu
lation of similar age and sex.
Wychulis et al23 reported on the ex
perience of 1318 patients treated for
achalasia between 1935 and 1967 at the
Mayo Clinic. Carcinoma developed in
seven patients for an incidence of 41 per
100 000 population. Wychulis et al con
sidered this rate to be elevated com
pared with the general US population,
but the rate estimates were not stan
dardized and there were no comparison
rates from the Mayo Clinic candidate
population. Chuong et al11 conducted a
prospective study of 100 patients with
achalasia in Connecticut. They found no
cases of esophageal cancer and concluded
that their negative results cast doubt on
the association. The number of patients
in the study and the follow-up (mean, 77
months) clearly not sufficient to
were
detect increased risk. A recent pro
an
spective study from the Netherlands re
ported the experience of a cohort of 195
patients with achalasia treated between
1973 and 1988.8 Three cases of cancer
developed during 874 person-years. The
incidence of cancer (3.4 per 1000) was
33-fold higher than that expected in the
population. A significantly increased risk
also was reported in a Danish study of
146 patients from 1949 to 1984.9 Both
the Dutch and Danish studies were hos
pital based rather than population based.
It is not certain whether the patients
served by referral hospitals can be com
pared with the general population. It is
also not clear whether prevalent cases
were excluded. A recent study in Italy
evaluated 244 patients with achalasia.10
The incidence of cancer in this group
was 18.6 per 100 000 population, a figure
that was not statistically different from
that expected. After excluding the first
year of observation, the incidence in our
study was 146.7 per 100 000 population.
Sonnenberg et al25 analyzed the records
of 15 000 patients with achalasia aged 65
years and older from Medicare hospital
discharge data files. Patients with acha
lasia also were more likely than the gen
eral Medicare population to have a di
agnosis of malignant neoplasm of the
esophagus (odds ratio, 6.4; 95% CI, 3.8
to 10.7). Comparison with hospitalized
patients may have biased the study to
ward the null.
To properly address cancer risk in
achalasia, certain methodological stan
dards are necessary. (1) All cases within
a defined geographic region should be
identified to prevent selection or refer
ral bias that might develop if cases were
chosen from a single hospital. (2) Pa
tients who have cancer at the start or
within a short time (eg, 1 year) should
be excluded to be certain that the acha
lasia is not secondary to the cancer. (3)
A person-years approach should be used
to calculate incidence. Patients should
contribute follow-up time until they de
velop cancer, die, or are censored by
migration or the ending date of the study.
(4) The study should have large enough

Downloaded From: http://jama.jamanetwork.com/ by a University of Pennsylvania User on 11/10/2015

 numbers to provide for adequate sta
tistical power. (5) There should be a suf
ficiently long follow-up to permit cancer
to develop. (6) There must be a suitable
comparison population to compare rates.
Of the published studies, ours is the only
one that meets these standards.
There are certain limitations to our
study. We have monitored patients from
the date of having their diagnosis regis
tered for the first time, but we do not
have information on the date of symptom
onset. Patients may go for many years
before seeking medical attention. It is
logical to presume that there is a defined
latency period between the onset ofacha
lasia and the development of cancer. Be
cause the SIRs in our study were stable
after the first year, one must conclude
that individuals in each ofthe subsequent
latency categories had their disease for
varying but comparable intervals. We
also have no information on the type of
treatment that patients received for acha
lasia. Esophageal dilation and stasis have
been thought to be responsible for can
cer. Perhaps patients who have under
gone effective surgical or balloon dilation
therapy do not have the same risks as
untreated patients. Similarly, we have
no information on the caliber of the
esophagus or on other risk factors for
esophageal cancer, such as cigarette
smoking. There was a lower risk of lung
cancer among patients with achalasia
compared with the reference population
(SIR, 0.7), which indirectly suggests that
smoking was not more common in the
achalasia cohort. The increased risks of
primary liver cancer (SIR, 2.3) and pan
creatic cancer (SIR, 1.9) imply that con-
References
1. Fagge CH. Case of simple stenosis of oesopha-
gus followed by epithelioma. Guy's Hosp Rep. 1872;
17:43.
2. Bolivar JC, Herendeen TL. Carcinoma of the
esophagus in achalasia. Ann Thorac Surg. 1970;10:
81-89.
3. Carter R, Brewer LA. Achalasia and esophageal
carcinoma: studies in early diagnosis for improved
surgical management. Am J Surg. 1975;130:114\x=req-\
120.
4. Frank MS, Brandt LJ, Haas K, Parker JG. Ma-
lignant esophagopulmonary fistula complicating
achalasia. Am J Gastroenterol. 1979;71:206-209.
5. Hankins JR, McLaughlin JS. The association of
carcinoma of the esophagus with achalasia. J Tho-
rac Cardiovasc Surg. 1975;69:355-360.
6. Lortat-Jacob JL, Richard CA, Fekete F, Testart
J. Cardiospasm and esophageal carcinoma: report
of 24 cases. Surgery. 1969;66:969-975.
7. Pierce WS, MacVaugh H III, Johnson J. Carci-
noma ofthe esophagus arising in patients with acha-
lasia of the cardia. J Thorac Cardiovasc Surg. 1970;
59:335-339.
8. Meijssen MA, Tilanus HW, van Blankenstein M,
Hop WCJ, Ong GL. Achalasia complicated by oe-
sophageal squamous cell carcinoma: a prospective
study in 195 patients. Gut. 1992;33:155-158.
9. Aggestrup S, Holm JC, S\l=o/\rensenHR. Does acha-
lasia predispose to cancer of the esophagus? Chest.
1992;102:1013-1016.
Downloaded From: http://jama.jamanetwork.com/ by a University of Pennsylvania User on 11/10/2015
founding by alcohol consumption, al
though unlikely, cannot be definitely re
futed. The follow-up time in this study
was limited, although some patients were
monitored for as many as 24 years. As
the cohort continues to grow and to age,
we will be able to make even more pre
cise estimates of risk. The fact that the
SIRs were stable after the first year sug
gests that the overall risk estimates will
not change appreciably as the cohort ages.
Finally, cancers of the gastric cardia were
not included in the risk estimates. It is
possible that some were adenocarinomas
that actually originated in the esopha
gus. This type of misclassification could
lead to an underestimate of esophageal
cancer risk from achalasia.
Esophageal cancer was a rare event in
both the cohort and the general popula
tion. Despite the relatively large size of
the achalasia cohort, the study had lim
ited power to assess increased risk in
certain subgroups. For example, only two
cancers were observed between years 1
and 24 in women with achalasia. Thus,
although the risk is increased eightfold
(SIR, 8.4), the 95% CI is broad (0.95 to
30.5) and includes the null value. In men,
the 95% CIs are also wide, but all of the
CIs comfortably exclude 1.0.
The information from this study can be
used to inform patients about their risk.
The data can also be used to generate
statistics that can be factored into the
decision-making process for surveillance.
Although previous evidence that sup
ported an increased risk for cancer was
incomplete or flawed, a number of au
thors have recommended periodic sur
veillance of patients with achalasia.13,14,22'26
10. Peracchia A, Segalin A, Bardini R, Ruol A,
Bonavina L, Baessato M. Esophageal carcinoma
and achalasia: prevalence, incidence, and results of
treatment. Hepatogastroenterology. 1991;38:514-516.
11. Chuong JJH, DuBovic S, McCallum RW. Acha-
lasia as a risk factor for esophageal carcinoma: a
reappraisal. Dig Dis Sci. 1984;29:1105-1108.
12. The role of endoscopy in the surveillance of
premalignant conditions of the upper gastrointes-
tinal tract: guidelines for clinical application. Gas-
trointest Endosc. 1988;34(suppl 3):18S-20S.
13. Codini RT, Levin B. Management of premalig-
nant lesions of the gastrointestinal tract. Practical
Gastroenterol. November/December 1981:6-10.
14. Heiss FW, Tarshis A, Ellis FH. Carcinoma as-
sociated with achalasia: occurrence 23 years after
esophagomyotomy. Dig Dis Sci. 1984;29:1066-1069.
15. Naess\l=e'\nT, Parker R, Persson I, Zack M, Adami
HO. Time trends in incidence rates of first hip
fracture in the Uppsala health care region, Sweden,
1965-1983. Am J Epidemiol. 1989;130:289-299.
16. The National Board of Health and Welfare.
Patient Register for Inpatient Somatic Care 1964\x=req-\
1983: A Quality Study [in Swedish]. Stockholm,
Sweden: National Board of Health and Welfare;
1991:20.
17. Lunde AS. The person number system of Swe-
den, Norway, Denmark, and Israel. Vital and Health
Statistics, Series 2: Data Evaluation and Methods
Research. Washington, DC: US Government Print-
In the current study, we have shown that
406 endoscopie procedures are needed to
detect a single cancer in men and 2200
are needed in women. This assumes an
nual examinations. If the interval were
less frequent than yearly, the number of
examinations would be fewer. Similarly,
if the examinations were postponed for
some interval after diagnosis, the num
ber of examinations would decrease. In
the current study, the cancer risk was
significantly elevated within 1 year of di
agnosis, although patients most certainly
had symptoms for longer than 1 year. It
is also important to recognize that there
is sparse evidence to suggest that sur
veillance will actually alter the natural
history of cancer. The role of endoscopy
for those treated late in their disease is
not known.12 We do not know if prema-
lignant changes in these patients can be
used to identify those at the highest risk.
We also do not know whether it would be
possible to detect early cancer in a di
lated, tortuous, debris-filled esophagus.
This study has demonstrated that pa
tients with achalasia are at substantially
increased risk of esophageal cancer com
pared with the general population. The
actual number of cancers that develop is
still small, and the benefits of surveil
lance remain unproven. Future studies
are needed to determine whether spe
cific patient characteristics can be used
to identify those at highest risk and to
possibly intervene to prevent deaths
from esophageal cancer.
This work was supported in part by a grants from
the American Society for Gastrointestinal Endos
copy and the National Institutes of Health (P30
DK34987, T32 DK 07634, and CP85636).
ing Office; 1980:5-11. DHHS publication PHS 80\x=req-\
1358.
18. Statistics Sweden. Causes of Death in Sweden.
Stockholm, Sweden: Liber; 1965-1984. Annual pub-
lications.
19. The Cancer Registry. Cancer incidence in Swe-
den 1958-1984. Stockholm: Swedish Board of Health
and Welfare; 1960-1988. Annual publication.
20. Tucker HJ, Snape WJ, Cohen S. Achalasia sec-
ondary to carcinoma: manometric and clinical fea-
tures. Ann Intern Med. 1978;89:315-318.
21. Camara-Lopes LH. Carcinoma of the esopha-
gus as a complication of megaesophagus: an analy-
sis of seven cases. Am J Dig Dis. 1981;6:742-756.
22. Just-Viera JO, Haight C. Achalasia and carci-
noma of the esophagus. Surg Gynecol Obstet. 1976;
128:1081-1095.
23. Wychulis AR, Woolam GL, Andersen HA, Ellis
FH. Achalasia and cancer of the esophagus. JAMA.
1971;215:1638-1641.
24. Dent TL, Kukor JS, Buinewicz BR. Endoscopic
screening and surveillance for gastrointestinal ma-
lignancy. Surg Clin North Am. 1989;69:1205-1225.
25. Sonnenberg A, Massey BT, McCarty DJ,
Jacobsen SJ. Epidemiology of hospitalization for
achalasia in the United States. Dig Dis Sci. 1993;
38:233-244.
26. Seliger G, Lee T, Schwartz S. Carcinoma of the
proximal esophagus: a complication of long-stand-
ing achalasia. Am J Gastroenterol. 1972;57:20-25.