The epidemiology of upper gastrointestinal bleeding

Crooks, Colin J. (2013) The epidemiology of upper gastrointestinal bleeding. PhD thesis, University of Nottingham.

PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (4MB) | Preview



There have been many conflicting changes in the prevalence of the risk factors for upper gastrointestinal bleeding and therefore it is not clear what the current trends in mortality or incidence are, nor which factors are important in driving these trends. As populations in many countries are ageing with an increasing burden of co-morbidity, this thesis investigates whether the relationship between non gastrointestinal co-morbidity and upper gastrointestinal bleeding might be an explanation for current trends. I hypothesised that non gastrointestinal co-morbidity was responsible for a large proportion of bleeds in the population and the deaths that occur following a bleed.


Large scale routine population based data records were used to assess the current incidence and mortality trends of upper gastrointestinal bleeding in England, as well as more in depth studies of predictors of its occurrence and subsequent mortality. The databases were examined and compared to external sources to assess their representativeness, and methods for defining cases in linked primary and secondary care were developed. The specific questions addressed in the studies were:

1. What are the current trends and variations in occurrence of upper gastrointestinal bleeding? Incidence rates and adjusted incidence rate ratios were calculated by quintiles of socioeconomic status, age group, sex, region, and calendar year.

2. Has there been an improvement in 30 day mortality following upper gastrointestinal bleeding? A nested case control study using Hospital Episodes Statistics from England 1999-2007 examined mortality trends by age, sex, co-morbidity and type of bleed.

3. Does non gastrointestinal co-morbidity predict upper gastrointestinal bleeding? A matched nested case control study used the linked Hospital Episodes Statistics and General Practice Research Database to examine non gastrointestinal co-morbidity as a risk factor adjusted for other known risk factors for bleeding. Sequential population attributable fractions were calculated to estimate what each risk factor contributed to the disease burden.

4. What are the excess causes of death following upper gastrointestinal bleeding? Causes of death by ICD 10 category were extracted following a bleed from the linked Office for National Statistics death register. Crude mortality rates and excess cumulative incidence functions were calculated; the latter adjusted for the competing risks between different causes of death.


1. A higher incidence of upper gastrointestinal bleeding was observed in the north of England, but this variation was dwarfed by the variation associated with deprivation. Areas of greater deprivation had 2-3 fold higher rates of hospitalisation for upper gastrointestinal bleeding than areas of less deprivation suggesting that strong modifiable risk factors exist.

2. Over the last decade there was a 20% improvement in 28 day mortality following upper gastrointestinal bleeding, and those admitted with bleeding were increasingly older and had more co-morbidity.

3. A combined measure of non gastrointestinal co-morbidity was a significant independent predictor of upper gastrointestinal bleeding and explained a greater proportion of the burden of bleeding (19%) than any other risk factor in the population, including medications such as aspirin and NSAIDs.

4. More than half the absolute excess risk of death was due to co-morbidity not related to the upper gastrointestinal tract.


Non gastrointestinal co-morbidity both strongly predicts an event of upper gastrointestinal bleeding, and is responsible for a large proportion of the subsequent long term mortality. The magnitude of the association in the population explains both why its incidence had not decreased, and why the improvements in mortality were observed irrespective of endoscopic management or bleed type. Furthermore a bleed can be an indicator for a re-assessment of the severity of co-existing non gastrointestinal morbidity.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Card, T.R.
West, J.
Subjects: W Medicine and related subjects (NLM Classification) > WI Digestive system
Faculties/Schools: UK Campuses > Faculty of Medicine and Health Sciences > School of Medicine
Item ID: 13394
Depositing User: EP, Services
Date Deposited: 08 Oct 2013 12:15
Last Modified: 15 Dec 2017 11:50

Actions (Archive Staff Only)

Edit View Edit View