37 by cuiliqing


									Epidemiology. 1998;9:452-456.


Luis Alberto García Rodríguez1 , Ana Ruigómez1 , Göran Hasselgren2,3,
Mari-Ann Wallander2,4, Saga Johansson2,5.

1 Centro Español de Investigación Farmacoepidemiológica (CEIFE).
Madrid, Spain.
2 Astra Hässle AB. Mölndal, Sweden.
3 Department of Surgery, Göteborg University, Sweden.
4 Department of Family Medicine, Uppsala University, Sweden.
5 Department of Medicine, Göteborg University, Sweden.

This study was supported by a grant of Astra Hässle.

Acknowledgments: We thank the staff at EPIC and the participating
general practitioners for their collaboration.

Correspondence to:
  Mari-Ann Wallander
  Astra Hässle AB. Epidemiology Dept.
  S-431 83 Mölndal, Sweden
  Tel. 46-31-776 12 53
  Fax. 46-31-776-37-33
  e-mail: lagarcia@ceife.es

Word count main text:           2,106
          abstract:             201

Abreviated Title: Peptic ulcer antecedents and mortality

Epidemiology. 1998;9:452-456.

We studied mortality related to peptic ulcer bleed (PUB) in a well defined

cohort of patients in the month following the episode of peptic ulcer

bleed. Our objective was to assess the contribution of peptic ulcer

antecedents and other predictive factors on the risk of dying. The study

cohort was formed by 1,020 patients with a hospitalization for an

episode of PUB between January 1991 and March 1994, identified in the

General Practice Research Database (GPRD) in the UK. Six hundred and

twenty three patients had no prior episode of peptic ulcer disease, 384

had peptic ulcer antecedents and in 13 information was not available.

Forty-five patients died (mortality rate: 4.4 per 100 person-months; 95%

CI: 3.3 - 5.9) within a month of the PUB. Patients with no peptic ulcer

antecedent faced a greater risk of dying than patients with antecedents

(RR=3.0; 95%CI: 1.2-7.1). Elderly patients, those undergoing surgery,

and current users of acid-suppressing drugs or non-steroidal anti-

inflammatory drugs (NSAIDs), all presented an increased mortality risk.

Patients presenting with their first ever episode of PUB carry a higher

case-fatality than those with previous episodes of peptic ulcer.

Keywords: Cohort study; mortality; peptic ulcer; bleeding;

      population-based study.

Epidemiology. 1998;9:452-456.

      There is a wide variation in mortality reported by different authors

for upper gastrointestinal bleeding. Estimates from several case series

of patients with gastrointestinal bleeding, including peptic ulcer and

other diseases, show a case-fatality between 2% and 14%1,2,3. Deaths

mainly occur among elderly people, or those with severe comorbidity 2.

Among the different underlying causes of upper gastrointestinal

bleeding, peptic ulcer disease is the single largest clinical entity,

accounting for 30-60% of all cases   2 ,4,5,6.   A substantial decline in

mortality rates from peptic ulcer has been reported in Europe since mid-

1980s. This trend has been attributed to therapeutic advances (including
the introduction of H2 -receptor antagonists) by some authors7,8. Other

reports did not show this decline among elderly patients9,10.

      Data on prognosis in patients with peptic ulcer bleed in recent

years are limited11,12. In addition, most of the published studies did not

include information on history of peptic ulcer. In one of these studies,

there was a suggestion that patients with a first episode of peptic ulcer

bleeding (PUB) presented a greater mortality risk than patients with

recurrent PUB1 1. To test this hypothesis, we conducted a study on the

short term mortality rate in a large population-based cohort of patients

with PUB in the UK. Our main goal was to study the mortality related to

peptic ulcer bleed in a well defined cohort of patients with an episode of

peptic ulcer bleed in the month following the episode of PUB. We

evaluated the role of peptic ulcer disease (PUD) antecedents and other

predictive factors on the risk of dying.

Epidemiology. 1998;9:452-456.


Data resource

     Approximately four million residents in England and Wales are

registered with general practitioners (GPs) who participate in the

General Practice Research Database (GPRD). This scheme comprises over

2,000 GPs who use computers in their offices for the purpose of

recording medical patient information in a standard manner. They have

agreed to provide the information anonymously to the Office of National

Statistics (ONS) and to allow the information to be used for research

projects. The information recorded includes demographics, medical

diagnoses, referrals to consultant and hospital, and a register of written

prescriptions. Prescriptions are automatically produced from the

computer and recorded on the patient's computerized file. A previous

study using this computerized data source has documented that over

90% of all referrals are entered on the GPs’ computers with a code that

reflects the clinical diagnosis13,14,15. An additional requirement of this

data resource is that the indication for any new course of therapy be

entered in the computer. In addition, the GP may record laboratory

results and other medical data in a free text comment field. A

modification of the OXMIS classification is used to code specific

diagnoses, and a drug dictionary based on the Prescription Pricing

Authority drug dictionary is used to record drugs.

Ascertainment of study cohort

      In the GPRD we identified patients aged 30 to 89 years old who

were admitted to hospital for an episode of upper gastrointestinal

bleeding between January 1991 and March 1994. Patients had to be

free from cancer, oesophageal varices, cirrhosis, Mallory-Weiss syndrome

and intestinal vascular pathologies. We reviewed manually the

computerized patient profiles of these patients to eliminate persons with

Epidemiology. 1998;9:452-456.
one of the following conditions: patients presenting only an episode of

haemorrhagic gastritis or duodenitis and patients with perforated peptic

ulcer. After revision, 1,020 patients with a diagnosis of peptic ulcer

bleeding involving the stomach or duodenum constituted our final study

cohort. Some of these patients were included in a previous epidemiologic

study evaluating the risk of upper gastrointestinal bleeding associated

with NSAID use16. We considered the first day of bleeding as the start

date of follow-up (index PUB) and followed each study member for a

period of one month. We identified all deaths and recorded the date. We

defined PUB related death as any death within a month of the episode

of PUB. We defined patients having prior episodes of peptic ulcer

disease, with or without bleed before start date, as patients with PUD

antecedents. The previous episode had to antecede the index PUB by

more than three months to be considered antecedent. For patients

where historical information was not complete on computer files

(N=772), we sent to the GPs a questionnaire requesting them to send a

copy of medical records and to determine whether the episode of index

PUB was the first ever or a recurrent one. Response rate was 88% and

patient confidentiality was preserved in all stages.

     Also, we used information from both the computer files and the

medical records for ascertaining the following variables. We classified the

site of bleeding using four categories: gastric and/or pyloric, duodenal,

multiple (both gastric and duodenal), and, for those with no specific

mention of site, as undiagnosed site. We recorded acid-suppressing
drug use before the index date (H2 receptor antagonists and proton

pump inhibitors) and categorized the exposure as non-use (when there

was not a single prescription before start date), current use (when the

most recent prescription was between day 1 and 60 before start date),

or past use (when the most recent prescription was more than 60 days

before start date). We categorized aspirin and NSAID exposure in a

Epidemiology. 1998;9:452-456.
similar way. We also recorded alcohol consumption and categorized

patients as abstainers/occasional drinkers when they imbibed less than

2 units of alcohol per week, light drinkers if they consumed 2 to 15 units

per week, and moderate to heavy drinkers when their consumption was

more than 16 units per week17. Body mass index (BMI) was also

recorded, as well as smoking. To ascertain comorbidity, we identified all

patients with a diagnosis before the start date of any of the following

groups of diseases: cerebrovascular, cardiovascular, respiratory

diseases and diabetes. We grouped all these diseases under the

variable, comorbidity. We created separate category for variables with

missing information.


     We analyzed PUB related mortality during the one month following

the episode of PUB, calculating the mortality rates and 95% confidence

intervals. We used life tables to compare survival for patients with or

without PUD antecedents. There was no live withdrawals from the life

table during the month of follow-up. We used Cox proportional

regression analysis to estimate relative risks (RRs) of mortality18.

Separate analyses were performed to see whether age, sex, site of PUB

and other factors such as smoking, BMI, alcohol consumption and drug

use were modifiers of the mortality risk. We reported estimates of RR

adjusted for age and sex.


     During the one month follow-up, 45 patients (4.4%) died in our

cohort of 1,020 cases of PUB; the remaining 975 were still alive at the

end of the month, as there was no loss to follow-up. Table 1 shows the

distribution of all studied potential risk factors according to antecedents

of PUD. Sixty two percent of the study cohort were males. The primary

source of hemorrhage was in the duodenum in 58% of the subjects.

There were 623 patients with no PUB antecedent, 384 with antecedents

Epidemiology. 1998;9:452-456.
and in 13 patients information was not available. Out of the 384 with

antecedents, 235 had a history of peptic ulcer without bleeding and the

remaining 149 had an earlier episode of bleeding peptic ulcer.

     During the first month, 45 patients died (mortality rate: 4.4 per 100

person-months; 95% CI: 3.3- 5.9). Of these deaths, 27% occurred

during the first day of admission. The probability of surviving for one

month was higher among patients with PUD antecedents (Figure 1). The

relative risk of dying in patients with no PUD antecedent was 3.0

(95%CI: 1.2 - 7.1) compared with patients with PUD antecedents (table

2). This increased risk was present in all age groups (data not shown).

The relative risk of dying among patients with no antecedent went up to

3.5 (95%CI: 1.4 - 8.8) when we adjusted for acid-supressing drug and

NSAID use.

     Patients with antecedents of peptic ulcer without bleeding had

approximately the same risk of dying compared with those with bleeding

peptic ulcer antecedents (RR=1.4; 95%CI: 0.3 - 7.7).

     Age was the most important predictor of mortality. Females

presented a higher mortality as well as patients taking acid-suppressing

drugs or NSAIDs at the time of the bleed. Case-fatality was the same

irrespective of the site of bleeding except for patients in whom the

primary site could not be ascertained (RR= 3.6; 95%CI: 1.4 - 8.8). Four

of the six fatal cases among patients with an undiagnosed site of

bleeding died during the first day. Another important predictor of

mortality was undergoing surgery (RR=6.0; 95%CI: 1.6 - 22.5). Patients

with comorbidity presented a similar mortality risk as those free of

comorbidity. None of the individual disease categories grouped under

comorbidity was associated with a relative risk greater than 1.5 (data

not shown).


Epidemiology. 1998;9:452-456.
     Patients with a first episode of PUB appear to have a greater risk

of dying than patients with recurrent episodes of peptic ulcer disease. In

this population, age was the major predictor of mortality, as in most

studies2,11,12,19,20. Although incidence of PUB was less in women than

men, case fatality among women was higher than in men2 , 2 0.

      Comparisons with previous studies are hindered by differences in

study populations.We only included patients with peptic ulcer

haemorrhage and not upper gastrointestinal bleeding due to other

causes. Twenty three percent and 15% in our study cohort had a history

of peptic ulcer without bleeding or with bleeding, respectively.

Corresponding figures in other studies were 17% and 10% in one

study21, and 22% and 10% in another22. Other studies analyzing only

antecedents of gastrointestinal bleeding reported estimates of

prevalence between 26% and 40%7 ,23.

      Haemorrhage from peptic ulcer is a serious condition associated

with a considerable mortality. Three follow-up studies among patients

with gastric ulcer bleeding reported a case-fatality ranging from 7% to

13% during the index admission1 9 , 2 0 , 2 2. More recent studies reported a

mortality rate at one month ranging from 5% to 7%5 ,24. Our estimate of

4.4% is near the lower extreme of that range.

      It has been reported that patients bleeding from a gastric ulcer are

at higher risk of dying than patients with a duodenal bleed1 1 ,25. We

found little difference; only patients with an undiagnosed site had a

greater mortality. As endoscopy is usually carried out within 24 hours of

admission, and more than half of deaths among patients with an

undiagnosed site occurred during the first day of hospitalization, one

could postulate that owing to the urgency of the clinical condition in this

subset of patients, either endoscopic examination was not performed, or

visualization of the site was impossible.

Epidemiology. 1998;9:452-456.
     Patients who had surgery carried a mortality risk six times greater

than for those not operated. Other studies with a larger proportion of

patients undergoing surgery found a similar increased risk between six

and eight fold11,19 related to surgery. Some authors found that the

presence of associated diseases, such as cardiovascular, chronic

respiratory, renal and liver failures, rheumatologic disorders and cancer

were major prognostic factors of all-cause mortality12,19,23,26,27. In our

study, we did not observed this relation. We only studied mortality in the

month subsequent to PUB, and we excluded patients with cancer, who

accounted for many of the excess deaths in a previous study1 2.

      In our data smoking was not associated with a major increased

risk of mortality after a PUB, similar to what has been reported by

others28. Patients with moderate or heavy alcohol consumption faced an

increased risk of mortality, but information on smoking, alcohol

consumption and body mass index was not available in a sizable

proportion of patients (20 to 37%).

      A study on risk factors for rebleeding and mortality of PUB patients

by Branicki showed that patients with no history of upper

gastrointestinal bleed had a higher mortality rate (7%) than patients

with a previous bleed (2%)1 1. In our study, we found the increase in

mortality to be three-fold in patients with no antecedent. Patients with

antecedents have survived at least one episode of peptic ulcer disease,

and are likely to receive more extensive careful medical management,

including preventive therapy and health care surveillance. They may also

seek earlier assistance. These factors, in addition to the different natural

history of peptic ulcer, may contribute to the better prognosis. In

addition, patients hospitalized for a PUB with history of peptic ulcer

disease may be more likely to be referred for early elective surgery1 9 , 2 7.

Epidemiology. 1998;9:452-456.
Table 1: Distribution of risk factors among those with and without
peptic ulcer disease (PUD) antecedents*
                              PUD antecedents† No PUD antecedents†
                              N=384            N= 623

30-59                       164    (42.7)    160   (25.7)
60-69                       85     (22.1)    160   (25.7)
70-79                       99     (25.8)    173   (27.8)
80-89                       36     (9.4)     130   (20.9)
Male                        273 (71.1)       349 (56.0)
Female                      111 (28.9)       274 (44.0)
Ulcer site
Gastric                     105    (27.3)    226   (36.3)
Duodenal                    250    (65.1)    339   (54.4)
Gastric & Duodenal          17     (4.4)     26    (4.2)
Undiagnosed site            12     (3.1)     32    (5.1)
Comorbidity status
Yes                         156 (43.3)       270 (40.6)
Surgical treatment
Yes                         14     (3.6)     53    (8.5)
Acid-suppressing drug use
Non use                     148 (38.5)       509 (81.7)
Current use                 102 (26.6)       6 4 (10.3)
Past use                    134 (34.9)       5 0 (8.0)
Aspirin use
Non use                     347 (90.4)       493 (79.1)
Current use                 2 9 (7.6)        121 (19.4)
Past use                    8   (2.1)        9   (1.4)
Non use                     311 (81.0)       375 (60.2)
Current use                 4 2 (10.9)       187 (30.0)
Past use                    3 1 (8.1)        6 1 (9.8)
Alcohol consumption
None/occasional             99     (25.8)    171 (27.4)
Light                       78     (20.3)    152 (24.4)
Moderate/Heavy              51     (13.3)    7 9 (12.7)
Body Mass Index
<25                         129 (33.6)       172 (27.6)
25-30                       8 8 (22.9)       164 (26.3)
>30                         3 3 (8.6)        6 5 (10.4)
Non smoker                  169 (44.0)       306 (49.1)
Smoker                      113 (29.4)       125 (20.1)
Ex-smoker                   3 6 (9.4)        5 2 (8.3)

Epidemiology. 1998;9:452-456.
* 13 individuals had missing information on PUD antecedents.
† Numbers presented may not sum up to total in some variables owing to missing
Table 2: Relative risk of mortality according to
PUB antecedents and other risk factors

                                RR    (95% CI)*

PUD antecedents
Y e s†                          1
No 3.0                          (1.2 - 7.1)
30-59†                          1
60-69                           2.8   (0.9- 9.1)
70-79                           4.2   (1.4 - 12.7)
80-89                           6.2   (2.0 - 19.1)
Male†                           1
Female                          2.0   (1.0 - 3.7)
Ulcer site
Gastric†                        1
Duodenal                        1.3   (0.6 - 2.5)
Gastric & Duodenal              0.7   (0.1 - 5.1)
Undiagnosed site                3.6   (1.4 - 8.8)
Surgical treatment
No†                             1
Yes                             6.0   (1.6 - 22.5)
Acid-suppressing drug use
Non use†                        1
Current use                     2.0   (1.0- 3.8)
Past use                        0.9   (0.3 - 2.3)
Non use†                        1
Current use                     2.0   (1.0 - 3.8)
Past use                        2.1   (0.8 - 5.3)
Alcohol consumption
None/occasional†                1
Light                           0.5   (0.2 - 1.7)
Moderate/Heavy                  2.1   (0.7 - 6.5)
Body Mass Index (BMI)
<25†                            1
25-30                           1.3   (0.5 - 3.2)
>30                             0     (0-0.9)
Non smoker†                     1
Smoker                          1.0   (0.4 - 2.5)
Ex-smoker                       1.4   (0.5 - 4.3)

Epidemiology. 1998;9:452-456.

* Adjusted RR derived from Cox regression models
including age, sex and the variable of interest.
† Reference category

    Epidemiology. 1998;9:452-456.


            .98                       PUD antecedents


                                            No PUD antecedents


                  0           7             14           21           28

                                     Follow-up days

Number at risk /deaths
PUD antecedents       384/5         379/0        379/0        379/1        378/0
No PUD antecedents 623/19           604/4        600/7        593/3        590/5

Figure 1: Kaplan-Meier survival plot29 for patients with and without PUD

          antecedents, and weekly life table totals

Epidemiology. 1998;9:452-456.
   Friedman LS, Martin P. The problem of gastrointestinal bleeding.
Gastroenterology. Clinics of North America 1993; 22(4): 717-721.
   Rockall TA, Logan RFA, Devlin HB, Northfield TC. Incidence of and mortality from
acute upper gastrointestinal haemorrhage in the United Kingdom. BMJ 1995;
   Hunt PS, Hansky J, Korman MG. Mortality in patients with haematemesis and
melena: a prospective study. BMJ 1979; 1: 1238-1240.
   Edge DP. Managenment of upper GI haemorrhage 1976-1987: A New Zealand
regional study. Gastroenterol Jpn 1991;26(Suppl 3):78-80.
   Mueller X, Rothenbuehler JM, Amery A, Meyer B, Harder F. Outcome of peptic
ulcer hemorrhage treated according to a defined approach. Worl J Surg 1994; 18:
   Longstreth GF. Epidemiology of hospitalization for acute upper gastrointestinal
hemorrhage: a population-based study. Am J Gastroenterol 1995; 90: 206-210.
   La Vecchia C, Lucchini F, Negri E, Reggi V, Levi F. The impact of therapeutic
improvements in reducing peptic ulcer mortality in Europe. Int J Epidemiol 1993;
   Penston JG, Crombie IK, Waugh NR, Wormsley KG. Trends in morbidity and
mortality from peptic ulcer disease:Tayside versus Scotland. Aliment Pharmacol
Ther 1993; 7: 429-42.
   Paimela H, Joutsi T, Kiviluoto T, Kivilaakso E. Recent trends in mortality from
peptic ulcer disease in Finland. Dig Dis Sci 1995;40: 631-5.
    Bloom BS, Fendrick AM, Ramsey SD. Changes in peptic ulcer and
gastritis/duodenitis in Great Britain, 1970-1985. J Clin Gastroenterol 1990; 12:
    Branicki FJ, Coleman SY, Fok PJ, Pritchett CJ, Fan S, Lai ECS, Mok FPT, Cheung
W, Lau PWK, Tuen HH, Lam S, Hui W, Ng MMT, Lam DKH, Tang APK, Wong J.
Bleeding peptic ulcer: a prospective evaluation of risk factors for rebleeding and
mortality. World J Surg 1990; 14: 262-269.
   Hudson N, Faulkner G, Smith SJ, Langman MJS, Hawkey CJ, Logan RFA. Late
mortality in elderly patients surviving acute peptic ulcer bleeding. Gut 1995;37(2):
    Jick H, Jick SS, Derby LE. Validation of information recorded on general
practitioner based computerised data resource in the United Kingdom. BMJ
1991;302: 766-768.
    Jick H, Terris BZ, Derby LE, Jick SS. Further validation of information recorded on
a general practitioner based computerised data resource in the United Kingdom.
Pharmacoepidemiol Drug Safety 1992;1:347-349.
    Van Staa J-P, Abenhaim L. The quality of information recorded on a UK
database of primary care records: a study of hospitalizations due to hypoglycemia
and other conditions. Pharmacoepidemiol Drug Safety 1994; 3;15-21.
    García Rodríguez LA, Jick H. Risk of upper gastrointestinal bleeding and
perforation associated with individual NSAIDs. Lancet 1994; 343: 769-772.
    Perry IJ, Wannamethee SG, Walker MK, Thomson AG, Whincup PH, Shaper AG.
Prospective study of risk factors for development of non-insulin dependent
diabetes in middle aged British men. BMJ 1995; 310: 560-4.
    Norusis MJ. SPSS Advanced statistics 6.1. SPSS Inc. Chicago IL: 1994, pp. 275-
    Coleman SY, Wong J, Pritchett CJ, Branicki FJ. Risk models for rebleeding and
postoperative mortality in bleeding gastric ulcer. Annals Royal College of Surgeons
of England 1991; 73:179-184.
    Smart HL, Langman MJS. Late outcome of bleeding gastric ulcers. Gut 1986;
    Fisher I, Madsen MR, Thomsen H, Høst V,Wara P. Peptic ulcer hemorrhage:
factors predisposing to recurrence. Scand J Gastroenterol 1994; 29:414-8.
    Røbæck-Madsen M, Fischer L, Thomsen H, Wara P. Late outcome of bleeding
gastric ulcer. Five to eight years' follow-up. Scand J Gastroenterol 1994; 29:983-

Epidemiology. 1998;9:452-456.
   Wilairatana S, Sriussadaporn, Tanphaiphat C. A review of 1338 patients with
acute upper gastrointestinal bleeding at Chulalongkorn University Hospital,
Bangkok. Gastroenterol Jpn 1991; 26(Suppl 3):58-61.
   Pérez Gutthann S, García Rodríguez LA, Raiford DS. Individual non-steroidal
anti-inflammatory drugs and other risk factors for upper gastrointestinal bleeding
and perforation. Epidemiology. 1997;8:18-24.
   Duggan JM. Is aggressive management of peptic ulcer justified by the data? J
Clin Gastroenterol 1993;13(2): 109-116.
   Rockall TA, Logan RF, Devlin HB, Northfield TC. Risk assessment after acute
upper gastrointestinal haemorrage. Gut 1996; 38(3):316-21.
   Nyhus LLM, Invited commentary to "Bleeding peptic ulcer: a prospective
evaluation of risk factors for rebleeding and mortality". World J Surg 1990; 14:
   Branicki FJ, Boey J, Fok PJ, Pritchett CJ, Fan S, Lai ECS, Mok FPT, Wong W, Lam
SK, Hui W, Ng MMT, Lok ASF, Lam DKH, Tang APK, Wong J. Bleeding duodenal
ulcer: a prospective evaluation of risk factors for rebleeding and mortality. Ann
Surg 1990; 211: 411-418.
   Norusis MJ. Kaplan-Meier survival analysis. In: Norusis MJ/SPSS Inc. SPSS
Advanced statistics 6.1. Chicago IL: SPSS Inc., 1994 : 275-290.


To top