(g,T,o,k)yo,gY,oNm,,I)4ed(,8,o,il)
LONG-TERM OUTCOME OF CHINESE PATIENTS WITH
ULCERATIVE COLITIS: EXPERIENCE WITH 259
PATIENTS OVER A 20-YEAR PERIOD
Ming-Jium SHIEH
Department of Gastroenterology (Director: Prof. Hiroshi OBATA)
Tokyo Women's Medical College (Received June 12, 1992)
The long-term clinical course of ulcerative colitis in Chinese was reviewed in a retrospective study of 259 patients, representing the total experience of a special clinic at National Taiwan University Hospital over a period of 20 years (1971--1990). There were 149 men and 110 women, and the age distribution was quite similar in both sexes, 70% of the patients having been first diagnosed at the age of 20-49.years. Status regarding the extent of colonic involvement revealed 151 patients (58.3%) with distal type, 46 patients (17.8%) with left type and 62 patients (23.9%) with total type of ulcerative colitis at their first visit. The commonest clinical course was chronic-intermittent (148/259==57%), followed by single-attack with subsequent remission (85/259=33%), and chronic-con-tinuous disease (26/259=10%). Within the first year of clinical observation proximal extension developed in 6% of the distal colitis patients and 7% of the left colitis patients. Only 10 (3.9%) of the 259 patients ultimately underwent surgical intervention, and the cumulative rate for surgery was 6.5% at 3 years, and 10% at both 10 years and 15 years. Eight deaths were documented, and the cumulative observed survival rate was exactly the same as the expected survival rate at all times during the first 15 years of elinical
follow-up. One female patient and one male patient with total colitis (2/62=3.2%)
developed synchronous colorectal cancer after being clinically observed for 16 years and 17 years, respectively. The cumulative risk of developing colorectal cancer in total coiitis patients was 2% (O-5%) at 16 years and 6% (O-13%) at 20 years, similar to figures quoted from Western reports. We concluded that while ulcerative colitis in Chinese patients pursues a peculiar but relatively mild course, it is accompanied by a considerable risk of colorectal cancer. Therefore, regularly scheduled colonoscopy for cancer surveillance is indicated in high-risk patients.
Introduction
Ulcerative colitis is a disease of the colonic
mucosa of unknown etiology. It is common in
Northern Europe and North America. Its
inci-dence is '5 to 15 new cases annually per 100,OOO population in these areasi}2). Ulcerative colitis
appears to be common among white populations but infrequent in ,such areas as Asia, South
America and Africa. On the other hand, the
annual incidence of ulcerative colitis in Japan has increased dramatically since the early 1960s from O.02 per 100,OOO population to a peak of O.47 in the
early 1970s, remaining stable at about O.4 per
100,OOO population since 19743). The actual inci-dence of ulcerative colitis in Taiwan is unknown, perhaps in part due to a less effectively disease registry system and a high incidence of infectious enterocolitis, resulting in the possibility that
clinical experience, however, cases of ulcerative colitis have been gradually increasing at referral centers in Taiwan, possible as a consequence of
improved diagnostic techniques and clinicians
awareness, or this may simply represent the
changing pattern of this disease entity in the
country.
Since ulcerative colitis was a relatively
uncom-mon disease in Taiwan, there have been few
reports regarding its clinical course. It is assumed, however, that a knowledge of the clinical course of ulcerative colitis will enable the clinician to treat the patient appropriately, to predict patient re-sponse to therapeutic modalities precisely, and to
help the patient cope with future problems. We
therefore retrospectively investigated patients with ulcerative colitis at National Taiwan Uni-versity Hospital (NTUH) and attempted to deter-mine the long-term outcome, comparing our find-'
ings with reports from Western countries and
Japan.
Materials and Methods
Patient Selection
The subjects of this retrospective study were
259 patients who visited National Taiwan Uni-versity Hospital (NTUH) during the period from
1971 to 1990 and who met the diagnostic criteria for ulcerative colitis, including typical clinicai
symptems, endoscopic appearance and patholog-ical findings, as described in the pioneering
studies conducted in England and Sweden. In this series, there were 208 patients seen at the time of their initial attack at NTUH and an additional 51 patients referred to us from other medical units during a relapse of long-standing colitis. Of the 51 relapse patients, 90% were referred within 2 years
of the onset'of symptoms. Since long-term
out-come, instead of short-term prognosis, was to be evaluated, both groups of patients were included iri this study. Patients who failed to survive the first presenting attack at NTUH (either the initial attack of colitis or the relapse of long-standing
colitis at the time of referral), patients who
underwent colectomy at the time of ,their first
visit at NTUH, and patients who were at first
diagnosed with ulcerative colitis and subsequent-ly found to have an other disease of the colon were excluded from this study.
Information'at the First Presenting Attack
"Age of onset" was defined as the year when the patient was first diagnosed instead of the year of
the onset of symptoms, since we considered the
latter to be unreliable. ,Routine laboratory tests
were performed to evaluate patient status, and
microscopic examination and culture of the stool were also carried out in all cases to rule out a specific infectious agent(s) as the cause(s) of the clinical symptoms. After an initial assessment of
colonoscopy in 1971 at NTUH, this examination
was performed in most patients after full bowel
preparation with the aim of ascertaining the
extent of disease and to take biopsies for patho-logical examination. Colon double-contrast studies
were only performed when colonoscopy was in-complete or refused by the patient. Extent of
disease could be classified into 3 types colono-scopically: a distal-type, a left-type and a total-type. The distal type is defined as mucosal change confined to the rectum. The left type is defined as
mucosal change beyond the rectum, extending
proximally no further than the hepatic flexure.
The total type is defined as mucosal change
beyond the hepatic flexure.
Follow-up Schedule
All patients were scheduled for clinical follow-up every three months, more frequently if symp-tomatic. As a rule, colonoscopy was recommended
to all patients every 6 months to one year to
evaluate disease activity endoscopically, to detect
the occurrence of proximal extension, and to
search for and identify malignant or premalignant lesions (severe dysplasia). As a means of cancer surveillance, multiple biopsies based on colono-scopic appearance were performed in addition. For purposes of this analysis, "precancer", "severe
dysplasia", and "high-grade dysplasia" are
re-garded as pathologically equivalent. Other specific
laboratory tests and imaging studies were also
performed, when indicated by the clinical
-tomatology. Potentially fatal local and systemic complications affecting the prognosis were
care-fully monitored. At the time of admission or surgery, the underlying etiology and extent of
disease were identified and the outcome described. The incidence of colorectal cancer according to the extent of disease at the first time of the examina-tion at NTUH and the duraexamina-tion of follow-up was calculated in order to provide accurate estimates
of the risk of developing colorectal cancer in
patients with ulcerative colitis. If patients failed to attend for two consecutive clinical reviews or one colonoscopic examination, telephone calls or letters were used to encourage re-attendance.
Final Reviews
Between January 1, 1971 and December 31,
1990, a total of 259 patients were admitted to the
follow-up program. The final status of the
pa-tients was reviewed on a day between January 1, 1991 and December 31, 1991. Since the outcome of patients lost to follow-up was treated the same as that of patient withdrawals; namely, it was as-sumed that the outcome of patients lost
to.follow-up was similar to that of patients with whom
contact was maintained and the data for the last
'
clinical'review were added to this study for
analysis. Although some biases may arise because of this assuMption, they should not invalidate our conclusions, since the number of losses was kept
to a minimum.
Because the extent of the disease has great
implications with respect to the severity of at-tacks and prognosis of ulcerative colitis, it has been adopted as a major parameter of our study to evaluate the long-term outcome of this disease.
The concept of "attack-years" could not be
em-ployed in our study, because there is no univer-sally accepted definition of an attack-year, not to mention "remission years", though it is common-ly used in the literature to describe the clinical course of ulcerative colitis. Some authors, for example, consider eveR one episode of bloody stool in a year clinically as constituting an attack-year, while others consider four or more bloody bowel movements a year necessary. In addition, in our
series a considerable proportion of patients ex-perienced only one attack with subsequent remis-sion lasting throughout the follow-up period.
Con-clusions drawn under such circumstances would
inevitably lead to misunderstandings.
Statistical Methods
The Chi-square test was used to compare the
proportion of each subgroup of extent of disease in relation to the status at the first presenting attack
at NTUH, and p values of <O.05 were considered
significant.
Because the subjects in our study did not enter
the study on a particular date but over an
ex-tended period of perhaps many years, there were difficulties in estimating the rate or risk of inter-ests. Hence, the "patient-year" concept was used to determine hospital admission rates and major complication rates for patients surviving the first
presenting attack at NTUH without surgery. A
patient-year is defined as one complete year of follow-up of an individual patient, and patients began to contribute to patient-year total at the end
of their first presenting attack at NTUH, and
.ceased to contribute when colectomy was resorted
to or they died. Such a definition offers two
advantages. It increases the material available for analysis, and permits allowances to be made for changes in the extent and severity of ulcerative
colitis4).
The clinical life-table method was used to
ana-lyze cumulative surgical and cumulative
survi-val rates. Moreover, the cumulative survisurvi-val rate was compared with the expected mortality rate of the population at large based on official vital
statistics5) with age and sex match for each calender year after entry into the study. The
Greenwood's formula was used to calculate
standard error and 95% confidence limits of the curve for comparison of the two survival curves. Since the number of patients followed for more
than 10 years was small, the Kaplan--Meier
life-table analysis was used to estimate the
cumu-lative risk of developing colorectal cancer and/or severe dysplasia. These risks were also expressed in terms of events per patient-year at different
times after the start of clinical observation.
Results
Population Studied
There were 149 males and 110 females eligible for this study, and the male : female sex ratio was 1.35. The age distribution of both sexes is shown in Table 1. The age distribution was quite similar in both sexes and yielded a uniform curve. Onset was at age 20 to 49 years in 70% of the patient, and at age 60 years or more in only 10%. The patient classification according to extent of disease at the time of the first presenting attack is shown in
Table 2. There were 151 patients (58.3%) with
distal colitis, 46 patients (17.8%) with left colitis, and 62 patients (23.9%) with total colitis at their
first visit at NTUH. When these patients were divided into two groups according to status at
their first visit (the initial attack of colitis or the relapse of long-standing colitis), obviously there tended to be a much lower proportion of patients with distal disease in the referred (relapse) group than in the initial attack group (25.5% vs. 66.3%).
Table 1 The age of onset and sex distributions of
patients with ulcerative colitis at NTUH
(1971-1990) ,
Ageofonset Male Female Total(%)
O-19 6 3 9(3.5) 20-29 27 25 52(20.1) 30-39 40 32 72(27.8) 40-49 36 24 60(23.2) 50-59 24 16 40(15.4) 60-69 14 8 22(8.5) 70-89 2 2 4(1.5) Total 149 110 259
This difference was statistically significant
(p<O.OOI). Based on the patients' clinical course, three distinct disease patterns could be observed:
A)
Achronicintermittentpatternwithremis-sion lasting from months to years, which was the
most common type and accounted for 57% (148/
259) of all cases.
B) A single-attack pattern with subsequent
remission lasting throughout the follow-up period, which accounted for 33% (85/259) of all cases.
C) A chronic continuous pattern without any
remission for years, which was the least common pattern, accounting for only 10% (26/259) of the cases.
Proximal Extension'
Extension to the proximal colon is summarized in Table 3. There were 9 cases of distal colitis
(9/151=6%) which spread proximally (7
develop-ing into left colitis and 2 into total colitis). Three cases of left colitis (3/46= 7%) also developed into total colitis. All of these proximal extensions occurred during the first year of observation and
were confirmed by colonoscopy. No further pro-ximal extension was detected colonoscopically
from the second year of follow-up to the end of the
review period.' No definite trigger factor was
observed in the cases extending proximally, and the therapeutic modalities were the same as in the other patients.
Measures of Morbidity
During the 20-year period, the 259 patients
contributed a total of 1,247 patient-years of
obser-vation. We used hospital admission and major
complication rates to assess ulcerative colitis morbidity. The major complications are listed in Table 4. They consisted of massive bleeding (16
Table 2 The relationship between extent of
ease and the status at the first presenting attack
Extent
ofdisease Totalno.
Initial
attack Relapticattack Distalcolitis Leftcolitis Totalcolitis 151(58.3%) 46(17.8%) 62(23.9%o) 138(66.30%)* 31(14.9%) 39(18.80/o) 13(25,50/o)' 15(29,40/o) 23(45.10/o) *p<O.OOI(distal type compared'with the other two types
combined)
Table 3 Proximal extenslon
Extent
ofdisease Firstvisit Finalreview Distalcolitis Leftcolitis Totalcolitis 151 46 62 142 50 67
-936-*Mstal colitis Left colitis Left Total Total colitis colitis colitis 7 cases 2 cases 3 casesTable4 Major complications of ulcerative
coiitis*
L Massivebleeding "'-･･･････････････････････････････+'･････+''16
2. Stricture ･-･･--･-･･---････････････････････････-･--･･･--･10
3. Perforation of the bowel ･･･････････････････････････････-･･･- 3
4. Toxic megacolon ･･'･･････････････････-････-･･･-･････-････ 2
5. Sepsis･･-･･･-･･-･･･i･･･････-･･････････i･･--+-･･･-････-････-････ 5 6. Dehydration with shock +････+････････････････････････'･･'･･ 1 7. Posttransfusion hepatitis ･･･････････････++･+･･･'･･'･･････････ 2
8. Primary sclerosing cholangitis ･･･････････････････'"''"' 1
Total episodes 40
"Exclusive of colon cancer and severe dysplasia.
events), stricture (10), perforation (3) and toxic megacolon (2). The systemic complications follow-ing or accompanyfollow-ing bowel inflammation included sepsis (5) and dehydration with shock (1). Another
two cases of posttransfusion hepatitis and one
case of primary sclerosing cholangitis were also diagnosed. None of the patients with distal colitis experienced any of the major complications men-tioned above. In all, there were 40 major complica-tion events in 33 patients (33/259=12.7%) during 1247 patient-years of follow-up, equal to an inci-' dence of 1 event per 31 patient-years. There were 31 hospital admissions (20 times for total colitis, 9 times for left colitis, and once for distal colitis), equal to an admission rate of once per 40 patient-years. It was our experience that patients with distal colitis seldom required hospital admission, either initially or during relapses.
Indications for Surgery and the Cumulative
Surgery Rate
Only 10 (3.9%) of the 259 patients ultimately
underwent surgery, including 9 patients with
total colitis and one patient with left colitis. The indications for surgery were cancer in 2 cases, massive bleeding in 3 cases, perforation in 3 cases and stricture with peritonitis in 2 cases. The two cancer-related operations were performed in the 16th and 17th year, respectively, of clinical
follow-up. One of the three patients with perforation
underwent surgical intervention in the 10th year
of follow-up, and the other 7 cases required
surgery within 3 years after the start of follow-up.
When these data, except the data for the 2 for
%
l2 l; g ii gig i ij t6･68･2
;
6
o 1 2 3 4 S 6 ÷ 8 9 10 11 12 13 14 15Years after entering the study
Fig. 1 Cumulativesurgeryratesincethestartofclinical follow-up
cancer patients, were analyzed from the time of entry into the study, using the clinical life-table method, the cumulative surgery rate was 6.5% at 3 years and 1096 at both 10 years and 15 years (Fig. 1). The reason why the'two cancer-related opera-tions were excluded from analysis was that these two operations occurred in the 16th and 17th year, respectivley, of clinical follow-up (details will be provided later), and the number of patients in our
series followed for more than 15 years was
ex-tremely small, meaning that the cumulative
surgery rate would be greatly affected by a single operation and the results of the analysis would therefore be very unreliable.
Causes of Death and the Cumulative Survival
Rate
At the end of review, 8 deaths had been docu-mented (8/259=3.1%) (7 patients with total colitis and 1 with left colitis), 5 of causes related to ulcerative colitis and 3 of causes other than colitis. The 5 ulcerative colitis-related causes of death consisted of cancer in 2 cases, bowel perforation with sepsis in 2 cases, and uncontrolled bleeding following colectomy in 1 case. All of these deaths, except the 2 colorectal cancer deaths in the 19th and 20th year respectively, of clinical follow-up, were randomly dispersed throughout the first 15 years of follow-up. No specific chronological pat-tern was found in either colitis-related deaths or colitis-unrelated deaths. We included both
o/co 1000 eg Ts S ,. ･g, g･ U e + Expected(age-andsex-matched) + Observed(95%confidence limits)
O123456789 10 11 12 13 14 '15
Years after entering the study
Fig. 2 Survical of patients with ulcerative colitis
compared with the age- and sex-matched background
population ts ,N 1.o o op Ng-n o op g- h
ov
g;
=o
o an,gios
oo
･B ℃ as ole
･S "v u g o.6 L----" l CancerCancer or severe dysplasia
l
L...-directly attributable to ulcerative colitis, as in the case of other authors. Based on these data, the cumulative survival rate in the first 15 yeats, excluding the two cancer-related deaths for the
same reason mentioned above, was calculated by
the clinical life-table method and compared with the expected survival rate. The results failed to
show any difference in survival between the
ulcerative colitis patients (95% confidence limits) and the background population at any time (Fig. 2). The two cancer-related deaths, however, sug-gested that malignancy is responsible for the late mortality in patients with ulcerative colitis.
Risks of Developing Cancer and/or Severe
Dysplasia
One female patient and one male patient with
total colitis (2/62 =3.2%) developed colorectal can-cer at age 38 and 42 years, respectively, after clinical observation for 16 years and 17 years,
respectively. The cancer surveillance programs were incomplete in these two patients, because they were reluctant to undergo regular colono-scopy, against advice. Both of thern underwent
colectomy after the diagnosis of cancer. Patholog-ical examination revealed multiple foci of poorly
differentiated adenocarcinoma in both surgical
specimens, one having 3 lesions and the other, 2 lesions. All 5 lesions were of the flat-ulcerative type grossly. The fact that both cancer patients
i
L.-.t
O s 10 15 20
Years after entering the study
Fig. 3 Proportion of patients with total colitis free of cancer and free of cancer or severe dysplasia since the' start of clinical observation
died of cachexia within 3 years after surgery,
suggests that they had advanced cancer when
diagnosed. Another three patients with total coli-tis (3/62 =4.8%) were demonstrated to have severe dysplasia by colonoscopy plus biopsy after clinical follow-up for 10, 12 and 15 years, respectively.
Prophylactic proctocolectomy was suggested but
refused by these patients. In short, no cancer or severe dysplasia (premalignant lesion) was diag-nosed less than 10 years after the start ef clinical follow-up. The cumulative risk of developing colo-rectal cancer in total colitis patients based on the Kaplan--Meier life-table analysis (with 95%
confi-dence limits) was 2% (O-5%) at 16 years and 6%
(O-13%) at 20 years, and the cumulative risk of
developing cancer or severe dysplasia was 2%
(O-5%) at 10 years, 9% (O-19%) at 15 years, and 17% (3-31%) at 20 years (Fig. 3). On the other,hand, when risk was calculated on the basis of
patient-years, the incidence shown in Table 5 was one
cancer per 130 patient-years, and one cancer or
severe dysplasia per 52 patient-years for the
second decade of follow-up.
Discussion
The present study revealed a higher risk of
ulcerative colitis between 20 and 49 years of age-approximately twice that of all other age groups
combined. Epidemiological studies in Western
countries6} have always stated that late occur-rence of this disease was more frequent in men,
while in the young the disease either occurred 938
-Table 5 Risk of developing colorectal carcinoma per patient-year'
No.ofpatientswith Riskperpatient-year
Durationof
follow-up
Totalof.
patlentyrs Carcinoma Severe
dysplasia Carcinoma Severedysplasia.orcarcmoma <10 10-20 402 260 o2 o3 o1/130 o1/52
'Only total colitis patients were included for analysis.
with equal frequency in both sexes or more often
in women. This impression, however, was not
confirmed in our study, which showed male pa-tients outnumbering female papa-tients in all age
groups. Generally, reports in the literature7)N9) have claimed that about half of the cases could be characterized as distal colitis and the other half as more extensive disease at the time the diagnosis was first confirmed. Our data appeared consistent with this trend, since distal colitis accounted for 58.3% of the patients in our study. However, it
was assumed to be higher among unselected
groups of patients, since most studies in the
literature have reported patients from hospitals and specialized clinics, which are not representa-tive of the general ulcerarepresenta-tive colitis population because of the higher rate of referral of severely ill patients and relatively reduced rate of referral of patients with extremely mild forms of the disease. In fact, in our study group the percentage of distal colitis in the initial attack group was significantly higher than in the referred (relapse) group. Hence, this bias due to partial patient selection must be taken into consideration when attempting to use the data from the literature to predict the clinical course and prognosis of ulcerative coiitis in a general population.
Based on the clinical course, the three distinct
disease patterns described above could be dis-criminated. Surprisingly, when compared with
the approximately 10% quoted in Western reports, as many as 33% of the patients in our series had only one attack with subsequent remission, pos-sibly reflecting an inability to rule out various possibilities of infectious colitis at the time of diagnosis, or this may represent a variant form of disease in Chinese patients, who in general are at
low risk for ulcerative colitis. In their well-known
reportiO) Edward and Truelove pointed out that
the longer the period of follow-up, the smaller the proportion of patients who had experienced only one attack and no recurrences. Whether, as time elapses, our series will exhibit the same trend merits further investigation. In several controiled trialsii) sulfasalazine and 5-aminosalicylic acid have been shown to have an action which reduces relapse rate from about 70% to about 20% per year
by an unknown mechanism. The widespread use
of sulfasalazine in ulcerative colitis started in 1973 at NTUH, and official reports will be pub-lished when adequate data are available.
In our series approximately 6% of the patients with distal colitis and 7% of those with left colitis developed proximal extension during the first year of clinical observation. This tendency to
proxi-mally spread was completely absent from the
second year of observation to the end of review period. When compared with the data of Farmeri2} and Hiwatashii3), who reported rates of 10% and 27.6%, respectively, our proximal extension rate was relatively low and did not increase with time.
Although the phenomenon of early proximal
ex-tension has been universally reported, most data in the literature show that the longer the follow-up period, the higher the proximal extension rate, underscoring the importance of the time factor. Why did our patient group present such a unique pattern of spreading? Was this just a variant form
of proximal extension or actually a maturation
process in the disease detected in the very early stage? Since we lack proper data to explain this peculiar manifestation, it remains an open
ques-tion.
Table 6 Major complications of ulcerative colitis (except colon cancer and severe dysplasia)
No. Perforation
Toxic
megacolon Massivebleeding Thrombo-embolisrn Stricture n o/o n o/o n o/o n o% n
%
Taiwan(1992) Uppsalai`)(1976" OxfordiO'(1963) 259 220 624 3720 L2 3.2 3.2 2510 O.8 2,2L6
16 821 6,2 3.6 3.4 o910 o4.1L6 10 oo 3.9 oowith insidious onset followed by a variable and protracted course and by many local and systemic complications. The incidence of major complica-tions in our series and two other series is listed and compared in Table 6. Interestingly, the major complications in our series differed considerably from those reported in OxfordiO} and Uppsalai4),
which showed a 2 to 4% complication rate for perforation, toxic megacolon, massive bleeding
and thromboembolism. Our own data, however,
showed a significantly higher incidence of mas-sive bleeding (16/259=6%), a relatively low inci-dence of perforation (3/259==1.2%) and toxic
mega-colon (2/259=O.8%) and no thromboembolism.
Vascular thromboembolism (including
pulmo-nary, cerebral and hepatic vessel thromboembol-ism) complicating ulcerative colitis is the result of increased blood coagulability (increased levels of factor, V, VIII and fibrinogen, accelerated throm-boplastin formation and decreased levels of anti-thrombin III, factor XIII and plasminogen'5)) in patients who are confined to bed as a result of severe illness. As is known, Chinese appear to be more resistant to such coagulopathy than Cauca-.
sians and our patients with uicerative colitis
seemed to have this inherited character. Massive
bleeding, on the other hand, was rather common
in our series, but usually it could be controlled or arrested by conservative treatments or by
colono-scopic hemostasisi6) and only one patient had
massive bleeding cited as a major cause of death. Interestingly, there were 10 patients (3.9%) exhibi-ting colonic stricture in a double-contrast study. Such stricture seldom occurs in Caucasian ulcera-tive colitis popuiation. Two of these 10 patients displayed spontaneous resolution of stricture in a double-contrast study later in the follow-up
per-iod. In general, patients with ulcerative colitis responded to home treatment. Hospital admission was required only for severe or fulminant attacks and life-threatening complications. There were 21 hospital admissions in our series and just as in the case of major complications, all of the patients admitted except one had total colitis or left colitis.
Thus, extent of disease and ulcerative colitis
morbidity were positively correlated. Similar re-sults were also observed in Uppsala where severe attacks affected patients with total colitis exclu-sively and did not affect any patients with left-sided or distal disease. As a matter of fact, all the operations and deaths in our series involved
pa-tients with total or left colitis.
According to Binder & Riis (1986), the colec-tomy rate a few years after the diagnosis takes a parallel course in patients with initial distal, left and total colon involvement, i.e., 1% a year. We had much better results; our cumulative rate for
surgery was 6.5% at 3 years and 10% at both 10
years and 15'years. Moreover, none of the patients with distal colitis underwent surgery during the follow-up period. It was unusual for patients with distal colitis in our series not to be controlled by
medical therapy, and surgical resolution was
rarely sought. The probable explanations, which
are also what we hope them to be, for the
relatively low surgical rate may be that our study is more recent, implying improved medical treat-ment, and that a high percentage of the patients in our series had only one attack. Nevertheless, is accepting these explanations optimistically with-out further assessment justified? Actually, when 'the indications for surgery in our 10 patients were thoroughly reviewed, it was found that, with the exception of the 2 patients with surgical
-tions because of cancer, the others were all for emergencies. In other words, none of our patients
underwent elective surgery and even the three patients with severe dysplasia refused
prophy-lactic proctocolectomy in spite of a strong
recom-mendation. Generally speaking, the "timing" of surgery is determined jointly by the
gastroen-terologist and surgeon and sometimes by the
patient. Although the indications for surgical
intervention, either elective or emergency, are now more clearly defined than was previously the case, most patients tended to postpone surgery, particularly in the young population, because of
dread of a permanent ileostomy. In addition,
gastroenterologists and surgeons have a psycho-logical obstacle to surgery, since ulcerative colitis may be mimicked by other types of colitis which are reversible, such as ischemic colitis,
pseudo-membranous colitis, campylobacter-associated
colitis or amebiasis of the colon, and they fear
making the grave error of submitting such pa-tients to proctocolectomy. Consequently, both
doctors and patients are prone to avoid surgical lnterventlon,
Our patient group displayed a lower total
mor-tality rate (3.1%) than those in Stockholm
countyi7} and Japani3}, for which total mortality rates of about 10% were reported. The experiences
in Scotland, Stockholm, Uppsala and Japan also
showed an accentuated mortality during the first few years after diagnosis, however, the subjects in all of these studies consisted of a select patient groups. Perhaps these patients who were the first to be closely observed were observed only because
they had more extensive disease and thus were
more inclined to die than ulcerative colitis pa-tients in the genera! population, or early excess mortality may actually be characteristic of ulcera-tive colitis. Although our patient group was de-finitely a select group, they did have comparably favourable survival figures, just as in the regional,
and therefore unselected patient group in the
study at Rochester!8). In our study, the cumulative observed survival rate, even with the 3
colitis-unrelated deaths included, was the same as the
expected survival rate, though the trend toward cancer deaths later in the course of the disease
was impressive. Generally, eur patient group
seemed to have a relatively mild form of disease. It is difficult to distinguish between different study
groups, however, because of the bias in metho-dology, patient selection, and completeness of foliow-up. Moreover, the results for long-term
prognosis, especially mortality, may reflect
dif-ferent factors such as new drugs and surgical methods developed during the study period. In
short, although ulcerative colitis mortality rates seem to vary from country to country, a general tendency to decline is evident in the data of the OMGE surveyi9). This may be because the disease itself has chaRged or because of the increasing use of sulfasalazine maintenance.
Agreement is virtually universal that ulcerative colitis leads to a significant increase in incidence of colorectal cancer. Data from a few studies in which emphasis was on defined areas and
popula-tions are summarized in Table 7. Most studies have provided evidence that ulcerative colitis
patients carry a higher risk of colorectal cancer than the general population and that this risk is concentrated in patients in whom the all or the
majority of the colon has been involved in the inflammatory process. There seems to be little
increase in risk of cancer over the risk in the general population up to 10 years after the onset of disease, but it does rise above the risk in the general population thereafter. As stated in the literature, risk rates for developing carcinoma in extensive colitis are estimated at between 5% and 10% at 20 years after the onset of symptoms, and at between 10% and 20% at 30 years after the onset of symptoms. Nevertheless, there continues to be differences in the results reported in studies on cancer incidence in ulcerative colitis, because
--most investigations are retrospective and differ m patient population, method of entry into the med-ical care system, and modes of patient assessment
and follow-up. Selection biases and different
methodology would certainly affect the results of outcome. For example, a, high colectomy rate had
Table 7 Incidence of colorectal cancer in ulcerative colitis Patients Cumulativeincidence(year-duration) n/Totaln 5 10 15 20 Taiwan(1992) London25)(1990) WestMidlands22'(1988) Oxford Stockholm Rochesteri8)(1987) Copenhagen2D)(1985) Uppsalai`)(1976) 2!62 22/401 38/823 3f179 7/783 71220 totalcolitis extensivecolitis extensivecolitis allpatients allpatients extensivecolitis ooooo2.5 ooO,71O,85,5 2* 33.41Ll8.4 657.22.51.4II,7 '16 year-duration
been claimed to account for a low incidence of
colon cancer in Copenhagen20);' and no cases of
cancer were observed in the excellent study by
Bonnevie2i), but this was not surprising, since
patients who had symptoms for 10 years in whom extensive colitis had been demonstrated were
considered as surgical candidates. In our series,
the cumulative risk of developing colorectal
cancer in total colitis patients was 2% (O-5%) at 16 years and 6% (O-13%) at 20 years. This incidence did not significantly differ from that reported in a
joint study of three cohorts in West Midland,
Oxford, and Stockholm22). Since none of our
pa-tients underwent prophylactic proctocolectomy during the follow-up period, our patient group
tended to show the true picture of cancer risk in ulcerative colitis. On the other hand, it should also be emphasized that in the surgical specimens from both our cancer patients multiple foci of
malig-nant transformation were found whose gross
appearance was flat-ulcerative. Although flat-type
colorectal cancers have not been uncommon in
the experience of expert Japanese endoscopists, these two cases were the only ones at our hospital. Moreover, the average age at the time of cancer
diagnosis in these two patients was 19 years
younger than in patients with non-colitis colorec-tal cancers (40 vs. 59) seen during the same review
period at NTUH23>. The findings that patients
with long-standing extensive colitis have an in-creased risk of coloreetal cancer, tend to develop
cancer early and exhibit a high incidence of
synchronism24), considered as a whOle, suggest the
possibility that the mechanism of carcinogenesis in ulcerative colitis patients'is different from that in non-colitis patients, and this has highlighted the importance of clinical approaches to cancer surveillance. Currently, colonoscopy is the prin-cipal procedure ultilized for surveillance, since colorectal cancer can arise in ulcerative colitis "de
novo" and be hard to identify in colon double-contrast studies. Premalignant lesions (severe
dysplasia) may also exist and required patholog-ical verification, and this has a great impact on
subsequent treatment policy. Whether
prophy-lactic proctocolectomy should be recommended for high-risk patients remains doubtful at present. We believe that as long as we are unable to alter the cancer risk in ulcerative colitis it will be
'
appropriate to ask whether prophylactic
procto-colectomy should be performed in patients who have had total colitis for more than 10 years. There is agreement, however, that colonoscopic
surveillance with multiple biopsies had modified the need for prophylactic proctocolectomy in these high-risk patients and enabled a more selective
use of proctocolectomy in patients with severe
dysplasia, or dysplasia, even when low-grade and
at multiple sites25)-27).
Hence, this retrospective study has shed light
on the characteristics of long-term outcome in
Chinese patients with ulcerative colitis. The
lit-eral meaning of "natural history" is that the
course of disease is free of diagnostic and
thera-peutic interference. This situation has seldom
existed, since the medical profession has always
tried to improve the course of disease, and has at times, because of the adverse effect of treatment,
made conditions worse. Our own findings
pre-sented here decidedly do not represent the natural history of ulcerative colitis but only our overall experience with our own selected population. The main problem of course, the same as others have faced, is that the patients selected had follow-up
periods of varying length. However, a
well-designed clinical study on well defined patient groups followed-up using clearly defined protocols
of treatment is the best tool for determining
future policy in treating patients. As a conse-quence, we would be able to improve their clinical course and prognosis by offering the right treat-ment at the right time.
Acknowledgements
First of all, I should like to acknowledge with
grati-tude the assistance I have received from Professor Dr. Hiroshi Obata for his excellent instruction and kindly
checking of this report in detail. I am especially indebted
to Professor Dr. Cheng-Yi Wang, Department of Internal Medicine, National Taiwan University Hospital, for his generous support, invaluable inspiration and long-term encouragement in this study. I am also grateful to Associate Professor Dr. Kou Nagasako for generously sharing his knowledge and providing advice. Lastly, I would like to express my sincere appreciation to my friends at the Department of Gastroenterology, Tokyo Women's Medical College, who have provided the crucial personal support, psychological and material, that makes academic study not only possible but rewarding.
References
I) Calkins BM, Mendeloff AI: Epidemiology of
matory bowel disease. Epidemiol Rev 8: 60-91, 1986
2) Nordenvall B, Brostom O, Burglund M et al:
Incidence of ulcerative colitis in Stockholm county 1955-1979. ScandJ Gastroenterol 20: 783-790, 1985 3) Utsunomiya T, Yokota A: Concept, epidemiology and etiology of inflammatory bowel disease. in Clinical
Inflammatory Bowel Disease (Shiratori T ed) pp 18-31,
Herusu Shuppan, Tokyo (1988) (in Japanese)
4) Watts JM, de Dombal FT, Watkinson G et al:
Long-term prognosis of ulcerative colitis. Br Med J 1:
1477-1453, 1966
5) Department of Health, the Executive Yuan:
Health and Vital Statistics Vol 2. Vi･tal Statistics, Republic of China (1990)
6) Binder V, Both H, Hansen PK et al: Incidence and
prevalence of ulcerative colitis and Crohn's disease in
the county of Copenhagen 1962-1978. Gastroenterology
83: 563-568, 1982 '
7) Morris T, RhodesJ: Incidence of ulcerative colitis inthe Cardiff region. Gut 25: 846:848, 1984
8) Berner J, Naer T: Ulcerative colitis and Crohn's
disease on Faroe Islands 1964-1983. Scand J terol 21, 188-192, 1986
9) Haug K, Schrumpf E, Barstad S et al: Epidemiology
of ulcerative colitis in Western Norway. Scand J troenterol 23: 517-522, 1988
10) Edwards FC, Truelove SC: The course and prognosis
of ulcerative colitis 1 and 2. Gut 4: 299-315, 1963
ll) Lauritsen K, Laursen LS, Bukhave K et al: Use of
colonic eicosanoid concentrations as predictors of lapse in ulcerative colitis: double blind placebo trolled study on sulphasalazine maintenance treatment.
Gut 29: 1316-1321, 1988
12) Farmer RG: Long-term prognosis for patients with
ulcerative proctosigmoiditis (ulcerative colitis confined to rectum and sigmoid colon). J CIin Gastroenterol 1:
47-50, 1979
13) Hiwatashi N, Yamazaki H, Kimura M et al:
ical course and longterm prQgnosis of Japanese patients with ulcerative colitis. Gastroenterol Jpn 26: 312-318,
1991
14) Brostom O: Prognosis of ulcerative colitis. Med Clin
North Am 74: 201-218, 1990
15) Kirsner JB: Inflammatory bowel disease Part II: ical and therapeuti'c aspects. Dis Mon 37: 673-746, 1991
16) Wang CY, Wong JM, Shieh MJ: Aggressive scopic approaches to lower intestinal bleeding.
trenterol Jpn 26 (Suppl 3): 125-128, 1991
17> Brostrom O, Monsen U, Nordenvall B et al:
Prognosis and mortality of ulcerative colitis in
holm county 1955-1979. Scand J Gastroenterol 22: 907-913, 1987
18) Stonnington CM, Philips SE Zinsmeister AR et
al: Prognosis of chronic ulcerative colitis in a
munity. Gut 28: 1261-1266, 1987
l9) Softley A, Clamp SE, Watkinson G et al: The
history of inflammatory bowel disease: Has there been a
change in the last 20 years? ScandJ Gastroenterol 23
(Suppl 144): 20-23, 1988
20) Hendriksen C, Kreiner S, Binder V: Long term
prQgnosis in ulcerative colitis-based on results from a regional patient group from the county of Copenhagen.
Gut 26: 158-163, 1985
21) Bonnevie O, Binder V, Anthonisen P et al: The
prognosis of ulcerative colitis. Scand J Gastroenterol 9:
81-91, 1974
22) Gyde SN, Prior P, Thompson H et al: Colorectal
cancer in ulcerative colitis: A cohort study of primary referrals from three centers. Gut 29: 206-217, 1988
23) Shieh MJ, Wong JM, Wang CY: Site distribution of
Chin J Gastr㏄nterol 7：116−121，1990 24）Ritchie J， Hawley P， Lennard・JonesJ：Prognosis of carcinoma in ulcerative colitis． Gut 22：752−755，1981 25）Fossard JBJ， Dixon MF：Colonoscopic surveillance in ulcerative colitis・dysplasia through the looking glass． Gut 30：285−292，1989 26）Lennard・Jones JE， Melville DM， Morson BC et a1：Precancer and cancer in extensive ulcerative colitis： finding among 401 patients over 22 years． Gut 31： 800−806，1990 27）Blackstone M， R董dell R， Rogers B et a1：Dysplasia− associated lesion or mass（DALM）det㏄ted by colono・ scopy in long−standing ulcerative colitis：an indication for col㏄tomy． Gastr㏄nterology 80：366−374，1981 中国人の潰瘍性大腸炎の長期予後に関する検討 一20年間259症例による経験一 隅京女子医科大学 消化学内科学教室（主任：小幡 裕教授） シヤ メイ キン
