Effects of Community Screening for Helicobacter pylori: 13-Year Follow-Up Evaluation of a Randomized Controlled Trial

Accepted Manuscript Effects of Community Screening for Helicobacter pylori: 13-Year Follow-up of a Randomized Controlled Trial (HEP-FYN) Maria Bomme Høgh, MD, Jane Møller Hansen, MD PhD, Mette WildnerChristensen, MD PhD, Jesper Hallas, DMSci, Ove B. Schaffalitzky de Muckadell, DMSci PII: DOI: Reference:

S1542-3565(17)30710-3 10.1016/j.cgh.2017.06.006 YJCGH 55288

To appear in: Clinical Gastroenterology and Hepatology Accepted Date: 6 June 2017 Please cite this article as: Høgh MB, Hansen JM, Wildner-Christensen M, Hallas J, Schaffalitzky de Muckadell OB, Effects of Community Screening for Helicobacter pylori: 13-Year Follow-up of a Randomized Controlled Trial (HEP-FYN), Clinical Gastroenterology and Hepatology (2017), doi: 10.1016/j.cgh.2017.06.006. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Effects of Community Screening for Helicobacter pylori: 13-Year Follow-up of a Randomized Controlled Trial (HEP-FYN)

Authors Maria Bomme Høgh, MD (Corresponding author)

RI PT

Community screening for Helicobacter pylori

SC

Department of Medical Gastroenterology, Odense University Hospital, Odense, Denmark; Institute of Clinical Research, University of Southern Denmark, Odense, Denmark

E-mail: [email protected]

M AN U

Phone: +45 282 93 973

Address: Department of Medical Gastroenterology, Odense University Hospital, Sdr. Bouleward 29, DK-5000 Odense C, Denmark

TE D

Jane Møller Hansen, MD PhD

Department of Medical Gastroenterology, Odense University Hospital, Odense, Denmark;

EP

Institute of Clinical Research, University of Southern Denmark, Odense, Denmark

AC C

Mette Wildner-Christensen, MD PhD Department of Medical Gastroenterology, Odense University Hospital, Odense, Denmark

Jesper Hallas, DMSci

Institute of Clinical Pharmacology, University of Southern Denmark, Odense, Denmark

Ove B. Schaffalitzky de Muckadell, DMSci

1

ACCEPTED MANUSCRIPT

Department of Medical Gastroenterology, Odense University Hospital, Odense, Denmark; Institute of Clinical Research, University of Southern Denmark, Odense, Denmark

Financial support

RI PT

The work was financially supported by the Region of Southern Denmark, Institute of Clinical Research University of Southern Denmark, The Research Board of Odense University Hospital, and private foundations: Aase and Ejnar Danielsens Foundation,

SC

Becket Fonden, and Helsefonden.

The funding had no role in study design, data collection, analysis or interpretation or in

Disclosures The authors have nothing to disclose.

TE D

Potential competing interests: None

M AN U

writing the report.

Guarantor for the article: Maria Bomme MD.

EP

Specific author contributions: MB, JMH, MWC, JH, and OBSM have provided substantial contributions in planning and conducting the study, interpreting data, drafting

AC C

manuscript, critical revision and have approved the final draft submitted.

2

ACCEPTED MANUSCRIPT

ABSTRACT

RI PT

Background & Aims: Helicobacter pylori (Hp) eradication improves the prognosis of peptic ulcer disease (PUD), dyspepsia, and possibly gastric cancer. Hp screening tests are accurate and eradication therapy is effective. Hp population screening seems attractive.

SC

The aim of this study was to evaluate the long-term effect of Hp population screening and eradication on dyspepsia prevalence and the incidence of PUD, and as secondary

M AN U

outcomes to assess the effect on health-care consumption and quality of life. Methods: At baseline in 1998−1999 20,011 individuals aged 40−65 years were randomized to Hp screening and eradication or a control group with no screening. Both groups received a questionnaire on dyspepsia and quality of life. Register data were

TE D

obtained for all randomized individuals.

Results: Baseline questionnaire response rate was 63%. Of the 5,749 screened, 1,007 (17.5%) individuals were Hp positive. Complete symptom data were obtained for 8,658

EP

(69%) individuals after 13 years. Dyspepsia prevalence decreased in both groups during follow-up, but multivariate analysis showed no effect of Hp screening and eradication

AC C

(adjusted OR=0.93 (0.82−1.04)); compared to usual care. Intention to treat and per protocol analysis of register data gave similar results. Hp screening did neither reduce the PUD incidence significantly (adjusted OR=0.88 (0.70−1.11)) nor did it have beneficial effect on health care consumption. Hp screening had no long-term effect on quality of life. Conclusion: This randomized clinical trial with 13-years follow-up, designed to give evidence on the effect of Hp population screening, showed no significant long-term effect when compared to usual care in this low prevalence area.

3

ACCEPTED MANUSCRIPT

www.ClinicalTrials.gov (identifier: NCT02001727).

Keywords: Helicobacter pylori screening; randomized controlled trial; dyspepsia; peptic

AC C

EP

TE D

M AN U

SC

RI PT

ulcer disease.

4

ACCEPTED MANUSCRIPT

BACKGROUND There is evidence that eradication of Helicobacter pylori (Hp) infection will improve this

RI PT

individuals’ prognosis with respect to peptic ulcer incidence 1, peptic ulcer recurrence 2, dyspepsia 3 and ulcer complications during NSAID therapy 2, and might reduce the subsequent risk of gastric cancer 4.

SC

Recurrence after Hp eradication or new infection in adulthood is rare 5, therefore screening is a one-time process for every birth cohort. 13C-urea Breath Test (UBT) is accurate,

M AN U

inexpensive and non-invasive and Hp can be permanently eradicated by a well-tolerated one-week therapy 6. From a public health perspective, it thus seems attractive and possibly cost-effective to actively trace Hp infected individuals at a population scale. HEP-FYN is the first study designed to show the long-term effect of Hp screening at

TE D

population scale 7. Additional register data on outcome and a large number of covariates were accessed for all randomized individuals, enabling us to evaluate the effect for all invited to screening and to assess the characteristics of nonresponders and thereby to

AC C

METHODS

EP

evaluate selection bias.

Study design and study population HEP-FYN is a prospective intervention study initiated in 1998−1999. At baseline, 20,011 individuals aged 40−65 years living in the county of Funen, Denmark were identified in the population register and allocated by a blind computerized randomization procedure to the Hp screening and treatment group or the control group (no screening). Randomization was

5

ACCEPTED MANUSCRIPT

done prior to invitation. Both groups received a questionnaire at inclusion, and after 1, 5 and 13 years. The intervention was an invitation to screening for Hp by an in-office blood test. Positive tests were confirmed by positive UBT. Hp positive individuals were offered triple

RI PT

eradication treatment (omeprazole 20 mg, clarithromycin 500 mg, and amoxicillin 1 g (or metronidazole 500 mg) all twice daily one week). In a random sample (n=200) UBT at least 4 weeks after eradication therapy showed a Hp eradication rate of 95% 7.

SC

Individuals who answered the 5-year questionnaire, alive and living in the county of Funen at the time of 13-year follow-up were sent a questionnaire. Second questionnaires were

M AN U

sent to nonresponders.

Outcomes

Our primary outcomes were prevalence proportion of dyspepsia at 13-year follow-up and the cumulative incidence of peptic ulcer disease (PUD). Secondary outcomes included

EP

Questionnaire data

TE D

prevalence of reflux, dyspepsia related health care consumption and quality of life.

The questionnaire included the Gastrointestinal Symptom Rating Scale (GSRS) 8 used to

AC C

measure gastrointestinal symptoms within the last week. Dyspepsia was defined as a score of ≥2 in the GSRS abdominal pain dimension corresponding to mild to severe discomfort 8. Gastro-esophageal Reflux Disease (GERD) was defined as a score ≥2 in the reflux dimension in the GSRS. The questionnaire covered smoking and average weekly alcohol consumption as well as dyspepsia-related sick leave days, general practitioner (GP) consultations, and use of

6

ACCEPTED MANUSCRIPT

ulcer drugs, antacids, aspirin and NSAID including over-the-counter (OTC) use during the previous month. Quality of life (QoL) was assessed by the highly validated Medical Outcome Study Short

RI PT

Form (SF-36) health survey 9.

Register data

All national registers can be linked using the unique personal identification number

assigned to all Danish residents at birth or upon immigration. Register data were accessed

SC

for all randomized individuals via Statistics Denmark 10.

M AN U

Data were obtained for two periods. The first period comprised the four years prior to index date and was used to evaluate responder vs. nonresponder differences. The second period was from index date to the 13-year follow-up date.

Data on all hospital contacts were obtained from the Danish National Patient Register (NPR) 11. Diagnoses were recorded according to the 8th and 10th revision of the WHO

TE D

International Classification of Diseases (ICD) 12. The first recorded PUD diagnosis verified by upper endoscopy or surgery was categorized as index ulcers and sub-categorized as

EP

complicated or uncomplicated. Patients with a diagnosis of gastro-esophageal cancer and esophagitis were identified. The number of upper endoscopies, hospital admissions and

AC C

outpatient visits with and without a related diagnosis of PUD were counted. Comorbidity is an independent risk factor of gastrointestinal bleeding 13. In order to correct for imbalances between groups, Charlson´s Comorbidity Index (CCI) was calculated based on diagnoses for all hospital contacts in the NPR 14. The CCI was modified to include ICD8 and ICD-10 codes 12 (supplementary table 1 and 2). Diagnosis of PUD was disregarded when computing CCI for the analysis of the outcome incident PUD. We defined levels of comorbidity as low (0), medium (1) and high (≥2).

7

ACCEPTED MANUSCRIPT

From the Danish National Prescription Registry data was obtained on consumption of reimbursable ulcer drugs, Hp eradication therapy, low dose acetylsalicylic acid (ASA) (≤150mg), high dose ASA (≥500 mg), and NSAID. High dose ASA was reported together with NSAID use. Drug utilization was analyzed using the defined daily dose (DDD) 15.

RI PT

Information on “highest completed education level” was obtained from the Populations Education Register. For simplification the 16 International Classification of Education (ICED97) 16 levels were recoded into four categories basic school education, secondary or

SC

short tertiary school education, long tertiary education and unknown (supplementary table 3).

M AN U

The Danish Socio-economic Classification: SOCIO is based on labor-market attachment 17

. SOCIO levels are employed, unemployed, retired, and unknown.

From the Danish Civil Register we obtained information on whether the person was living

Statistical analyses

TE D

with a spouse or not.

Assuming a 1-yr dyspepsia prevalence of 25% in the population, inclusion of 7,000

EP

individuals per group would be sufficient to demonstrate a reduction in dyspepsia of 5%, accepting type 1 and type 2 errors of 5%. Conservatively, we assumed only individuals

AC C

with peptic ulcer would have an effect of Hp treatment on dyspepsia. A significant reduction in incidence of peptic ulcer and ulcer complications was seen in the screened group compared to the unscreened group at 5-year follow-up. We expected this effect would be larger in an aging population with an expected increase in ASA and NSAID consumption. The sample size is too small to assess effects on gastric cancer incidence in this low prevalence region.

8

ACCEPTED MANUSCRIPT Completeness of recruitment at baseline To evaluate whether there was imbalance in baseline covariates for participating subjects due to selective participation, we analyzed the association of baseline participation and

RI PT

randomization group, age, sex, previous PUD, comorbidity, SOCIO, education group, and cohabitation in a logistic regression model including all randomized individuals.

SC

Main outcomes

The point prevalence of dyspepsia at the 13-year follow-up was analyzed for all

M AN U

questionnaire responders at this follow-up (complete case analysis). Analysis of incident PUD was based on register data, therefore both Intention-to-treat analysis (ITT) and per protocol (PP) analysis were possible. The association of these main outcomes and randomization group was explored using univariate and multivariate logistic regression. To account for imbalances between groups and outcomes categorical variables for age, sex,

TE D

ASA/NSAID, LDA, CCI, and education were included in a multivariate logistic regression models. Furthermore presence of dyspepsia at baseline or previous PUD was included as

EP

a covariate in the final model with the outcome dyspepsia and PUD, respectively.

AC C

Secondary Outcomes

Differences between randomization groups were compared using chi-squared test, and Mann-Whitney U-test when appropriate. From questionnaire data differences in GP consultation rate, number of dyspepsia related sick leave days as well as Inter-group differences in QoL scores at baseline and at the 13year follow-up and changes in QoL from baseline to end of follow-up were assessed. Both PP and ITT analyses of secondary outcomes based on register data were performed.

9

ACCEPTED MANUSCRIPT Ethical considerations The Regional Medical Ethics Committee of the Region of Southern Denmark approved the protocol (project-ID: S-VF-20110054-97/262).

RI PT

Data were handled according to the laws regulating private research registers and authorized by the Danish Data Protection Agency (project ID: 14/15606) 18. The study was registered at ClinicalTrials.gov (identifier: NCT02001727).

SC

All authors had access to the study data and reviewed and approved the final manuscript.

M AN U

RESULTS Study Population Flow

Of the 20,011 invited individuals at baseline 12,530 (63%) were enrolled. A total of 5,749 were screened for Hp at baseline; 1,007 (17.5%) were Hp positive. Eradication therapy was refused or not completed in 24 of these Hp positive individuals. A total of 9,641

TE D

individual were eligible and invited to the 13-year follow-up, and 8,658 (90%) of these participated, thus resulting in complete questionnaire data for 69% of the responders

EP

enrolled at baseline (Figure 1).

AC C

Completeness of recruitment at baseline At baseline there were 4,258 (43%) nonresponders in the screening group and 3,223 (32%) in the control group. The mortality was significantly lower among baseline responders compared to nonresponders. During the 13 years 1,238 (9.9%) baseline responders vs. 1,189 (15.9%) nonresponders had died resulting in a risk ratio (RR) of 0.62 (95%CI: 0.58−0.67).

10

ACCEPTED MANUSCRIPT

At baseline more responders than nonresponders were female, had previous PUD, were living with a spouse, had a higher education level, and were less likely retired (table 1). Those above 55 year of age were more likely to participate than those in the youngest age

RI PT

group. There was no association between CCI score and baseline participation.

Characteristics of questionnaire responders at baseline and at the 13-year follow-up Comparison of baseline responders in the screening group and control group showed no

SC

difference in the distribution of age and sex (male 49.5% vs. female 48.6%, mean age 52.2 vs. 52.2). Those lost during follow-up were equally distributed in regard to sex and age

M AN U

between groups (male 50.5% vs. female 50.5%, mean age 52.6 vs. 52.4). Alcohol use was similar between groups and remained stable. The proportion currently smoking decreased in both groups during follow-up.

The self-reported proportion of daily users of LDA and ASA/NSAID was higher at 13-year follow-up than at baseline (LDA 5.9% to 17.4% and ASA/NSAID 3.2% to 4.1%). The

TE D

proportion of daily users of ulcer drugs was three times higher compared to baseline while

EP

the use OTC antacids decreased. These changes were similar between groups.

Long-term effect on dyspepsia prevalence

AC C

At baseline dyspepsia was significantly more prevalent in the screening group compared to the control group (24.8% vs. 21.0%, OR=1.26 (1.16−1.37)) (table 2) but not at the follow-up points. The prevalence of dyspepsia decreased slightly in both groups to 18.9% in the intervention group and 18.7% in the control group at the 13-year follow-up (Table 2). Of the responders with dyspepsia at the 13-year follow-up, 48.5% had also reported dyspepsia at baseline. Complete case analysis of dyspepsia prevalence at the 13-year follow-up using logistic

11

ACCEPTED MANUSCRIPT

regression showed no effect of Hp screening and eradication with an adjusted OR=0.93 (0.82−1.04) (Supplementary table 4). Those with dyspepsia at baseline receiving Hp eradication (n=172) did not have a significant lower risk of dyspepsia at the 13-year follow-up compared to the Hp negative

RI PT

(n=764) with baseline dyspepsia (RR 1.03 (95% CI: 0.89−1.19)).

Long-term effect on PUD incidence

SC

During follow-up a total of 316 incident PUD cases occurred among the 12,530 responders enrolled at baseline (135 screened group vs. 181 control group) (table 3). The PUD

M AN U

incidence rate was 1.9 vs. 2.2 per 1,000 persons per year, in the screening group and control group, respectively. The incidence rate ratio (IRR) was 0.87 (0.69−1.10). Univariate and multivariate logistic regression analysis showed no effect of randomization on incidence of PUD, neither in the ITT (OR=1.00 (0.80−1.13)) nor in the PP analysis (OR=0.88 (0.70−1.11)) (Supplementary table 5).

TE D

During the 13-year follow-up period 253 incident peptic ulcers occurred in nonresponders (n=7,481) − an incidence rate of 2.8 per 1,000 persons per year. The PUD incidence rate

EP

was significantly lower among baseline responders compared to nonresponders (IRR=0.72 (0.61−0.86)). The number of incident complicated peptic ulcers in the

AC C

responders enrolled was 45 in the screened group vs. 59 in the control group (table 3). The incidence of complicated peptic ulcers was significantly lower in individuals enrolled compared to nonresponders (IRR=0.57 (0.43−0.75)).

Gastro-esophageal reflux disease Analysis of baseline questionnaire data showed that GERD was significantly more prevalent in the screening group than in the control group (27.6% vs.23.7%, RR=1.16

12

ACCEPTED MANUSCRIPT

(1.10−1.24)). There was no difference between groups in self-reported reflux symptoms (RR=1.06 (0.98−1.15)) (table 2) for those successfully followed for 13 years. However, there was a significantly higher number of hospital contacts with the diagnosis of esophagitis in the screened group when compared to the control group (PP analysis,

RI PT

RR=1.24 (1.07−1.44)) (table 3).

Cancer

SC

The total number of incident upper gastrointestinal cancers was 90 (0.5%) in all

randomized individuals and 45 (0.5%) in those enrolled and the distribution was equal

M AN U

between groups. The incidence of upper gastrointestinal cancers was significantly lower in enrolled vs. nonresponders (RR=0.57 (0.38−0.86)). When comparing only incident gastric cancers (exluding esophagel, cardia, and duodenal cancers) the difference was not

Health resource use

TE D

significant (responders 26 (0.2%) vs. nonresponders 20 (0.3%), RR=0.78 (0.43−1.39)).

There was no difference in self-reported sick leave days or GP consultations for dyspepsia

EP

(table 2).

Register data for the 13-year follow-up are displayed in table 3. There was no difference

AC C

between groups in the proportion of individuals who had a PUD-related hospital contact neither in ITT nor in the PP analysis (RR=0.94 and RR=0.87, respectively). The proportion of individuals who had at least one endoscopy during follow-up was similar between groups. The proportion of individuals who had more than one endoscopy was about 13% in both groups. The number of proton pump inhibitor (PPI) users and the mean use of PPI (mean DDD per user) were significantly higher in the screened group for those enrolled compared to the

13

ACCEPTED MANUSCRIPT

control group of those enrolled (PP analysis). The number of H2-antagonist and ASA/NSAID users was significantly higher in the control group, but there was no difference in mean DDD per user between groups. These differences in use of prescription

Quality of Life

RI PT

medication were not seen in the ITT analysis.

At the 13-year follow-up 7,837 (91%) of responders filled in the questionnaires assessing

SC

QoL, 87 questionnaires were excluded due to logical faults.

The SF-36 dimension and summery scores PH and MH all decreased during follow-up in

observed (table 2 & figure 2).

DISCUSSION

M AN U

both groups. No difference in mean dimension or summary scores between groups was

In this population based, randomized clinical trial with 13-years follow-up, screening for

TE D

Helicobacter pylori neither reduced the prevalence of dyspepsia, the incidence of peptic ulcer disease significantly nor did it have a beneficial effect on health care consumption or

EP

quality of life.

The population sample was randomized prior to invitation and therefore the study shows

AC C

the effect of screening on a population scale and can contribute to the evidence on systematic screening from a public health perspective. Two other randomized controlled studies with long-term follow-up invited a random population-sample 19, 20, and only those found Hp positive by screening were randomized (to eradication therapy or placebo) and followed. These studies answer a related but entirely different question: whether a Hp positive screening test should be followed by eradication therapy. They thus evaluate the effects of Hp eradication exclusively in Hp positives. By including the baseline population

14

ACCEPTED MANUSCRIPT

in our design, we also account for possibly negative effects of screening in HP-negative subjects and in non-participants.

The high response rate is a strength. Also, the prospective collection of data and the short

RI PT

recall period of GSRS and SF-36 reduce recall bias. Access to valid national registers on individual level for all randomized individuals enabled us to characterize non-responders. This reduces the influence of selection bias introduced

SC

by risk factors that can be measured in registries. Furthermore it permitted event detection independent of participation, reducing nonresponse bias and bias from loss to follow-up

M AN U

and thus improves the internal validity.

From year 2000 the diagnostic coding was used as the basis for payment of hospitals via the Diagnostic Related Group (DRG) system. This provided a strong economic incentive for comprehensive coding, and the coding can therefore be assumed to be complete 11. Inand out-patient health care is almost exclusively under national health care services and

TE D

the population coverage can therefore also be regarded as complete. A limitation in using administrative data for event detection is that data were not collected

EP

for research purposes. This can, however, also be considered a strength, as data were collected independently of the intervention and reflect current clinical practice. Variation in

AC C

coding practices between doctors, departments, or institutions and over time poses a risk of misclassification bias. The inclusion of a non-valid case would tend to dilute the ORs toward unity. Using diagnoses from hospital contacts as the basis of comorbidity classification may introduce misclassification as this measures only illnesses that requires hospital contact.

At the 5-year follow-up a modest reduction in the rate of dyspepsia, consultation rate, and

15

ACCEPTED MANUSCRIPT

sick leave days for dyspepsia, but no reduction in drug consumption was observed 21. A significant reduction in incidence of peptic ulcer and ulcer complications was seen in the screened group compared to the unscreened group at 5-year follow-up 21. We expected this effect would be larger in an aging population with an expected increase in ASA and

RI PT

NSAID consumption. However, at the 13-year follow-up only a non-significant trend toward a lower incidence of peptic ulcer and ulcer complication was found in the intervention group compared to the control group. Several factors may explain this.

SC

Baseline questionnaire data revealed a possible selection effect; those with dyspepsia or reflux symptoms were more prevalent in the screening group than in controls. The analysis

M AN U

for selection showed that non-responders were characterized by being male, retired, living alone, and having a low level of education.

Low socioeconomic status 22 and being a single male are known risk factors of PUD 1 and the Hp prevalence is higher in this group22. Therefore, if more individuals had participated, the effect of Hp screening might have been greater. Attrition as an explanation is further

TE D

supported by the observed higher incidence of PUD, cancer, and death among nonresponders. The problem with attrition is known from other screening programs; the

24

.

EP

population group with assumed highest risk does not participate in screening programs 23,

AC C

In Denmark the prevalence of Hp is low and this further dilutes the effect of eradication. Furthermore, during follow-up 5% of the control group were Hp tested and 3% received eradication therapy. This screening of the control group, probably as part of the Hp “test and treat strategy” that has been implemented in Denmark during the last decade, dilutes the effect of the intervention further. However, the extent is small and does not compromise the study´s ability to provide evidence of the effects of population screening.

16

ACCEPTED MANUSCRIPT

At least three large studies have shown the neither Hp presence nor Hp eradication increases prevalence of GERD 19, 25, 26. These results are confirmed in this long-term follow-up in regards to self-reported symptoms of reflux. However, analysis of register data showed a significantly higher number of persons with a hospital contact with the diagnosis

RI PT

of esophagitis in the screened group compared to the unscreened group (p=0.004). This could reflect the imbalance at baseline in proportions reporting reflux symptoms (27.6% in screened group vs. 23.7% in the control group).

SC

There was at significant higher use of PPI in the screened group during follow-up and this could have enhanced a positive effect of screening on dyspepsia prevalence and PUD

M AN U

incidence. The higher use of PPI in the screened group could be due to a higher prevalence of GERD.

Screening in itself can have led to concern and anxiety in the screened group and led to over-reporting of symptoms. However, 48.5% of the responders with dyspepsia at the 13year follow-up had reported dyspepsia at baseline.

TE D

The SF-36 dimension and summary scores all decreased during follow-up in both groups. In comparison with Danish Norm data the same decrease is observed with increasing age 27

EP

. This study showed no long-term effect of Hp screening on Qol. Our findings are in line

AC C

previous long-term studies 19.

Conclusion

This randomized clinical trial with 13-years follow-up was designed to give evidence on the effect of Hp screening at a population scale. It showed no significant long-term effect of population screening when compared to current clinical practice in a low prevalence area.

17

ACCEPTED MANUSCRIPT REFERENCES

7.

8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

20. 21. 22.

23.

RI PT

6.

SC

5.

M AN U

4.

TE D

3.

EP

2.

Rosenstock S, Jorgensen T, Bonnevie O, et al. Risk factors for peptic ulcer disease: a population based prospective cohort study comprising 2416 Danish adults. Gut 2003;52:186-93. Vergara M, Catalan M, Gisbert JP, et al. Meta-analysis: role of Helicobacter pylori eradication in the prevention of peptic ulcer in NSAID users. Aliment Pharmacol Ther 2005;21:1411-8. Moayyedi P, Soo S, Deeks J, et al. Systematic review and economic evaluation of Helicobacter pylori eradication treatment for non-ulcer dyspepsia. Dyspepsia Review Group. Bmj 2000;321:659-64. Wong BC, Lam SK, Wong WM, et al. Helicobacter pylori eradication to prevent gastric cancer in a high-risk region of China: a randomized controlled trial. Jama 2004;291:187-94. Xia HH, Talley NJ. Natural acquisition and spontaneous elimination of Helicobacter pylori infection: clinical implications. Am J Gastroenterol 1997;92:1780-7. Malfertheiner P, Megraud F, O'Morain CA, et al. Management of Helicobacter pylori infection--the Maastricht IV/ Florence Consensus Report. Gut 2012;61:646-64. Wildner-Christensen M, Moller Hansen J, Schaffalitzky De Muckadell OB. Rates of dyspepsia one year after Helicobacter pylori screening and eradication in a Danish population. Gastroenterology 2003;125:372-9. Svedlund J, Sjodin I, Dotevall G. GSRS-a clinical rating scale for gastrointestinal symptoms in patients with irritable bowel syndrome and peptic ulcer disease. Dig Dis Sci 1988;33:129-34. Ware JE, Jr., Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992;30:473-83. Statistics Denmark. The Danish system for access to micro data. Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Register. Scand J Public Health 2011;39:30-3. World Health Organization. International Classification of Diseases (ICD). Crooks CJ, West J, Card TR. Comorbidities affect risk of nonvariceal upper gastrointestinal bleeding. Gastroenterology 2013;144:1384-1393. e2. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987;40:373-83. Kildemoes HW, Sorensen HT, Hallas J. The Danish National Prescription Registry. Scand J Public Health 2011;39:38-41. UNESCO Institute for Statistics (UIS). The International Standard Classification of Education ICED 2011.: Montreal, Canada. 2012. Statistics Denmark. SOCIO - Danmarks Statistiks Socioøkonomiske Klassifikation 1997. Danish Data Protection Agency. The Act on Processing of Personal Data. Act No.429 of 31 May 2000, with amendments[Internet]. Lane JA, Murray LJ, Noble S, et al. Impact of Helicobacter pylori eradication on dyspepsia, health resource use, and quality of life in the Bristol helicobacter project: randomised controlled trial. Bmj 2006;332:199-204. Ford AC, Forman D, Bailey AG, et al. A community screening program for Helicobacter pylori saves money: 10-year follow-up of a randomized controlled trial. Gastroenterology 2005;129:1910-7. Hansen JM, Wildner-Christensen M, Hallas J, et al. Effect of a community screening for Helicobacter pylori: a 5-Yr follow-up study. Am J Gastroenterol 2008;103:1106-13. Rosenstock SJ, Jorgensen T, Bonnevie O, et al. Does Helicobacter pylori infection explain all socioeconomic differences in peptic ulcer incidence? Genetic and psychosocial markers for incident peptic ulcer disease in a large cohort of Danish adults. Scand J Gastroenterol 2004;39:823-9. Frederiksen B, Jørgensen T, Brasso K, et al. Socioeconomic position and participation in colorectal cancer screening. British journal of cancer 2010;103:1496-1501.

AC C

1.

18

RI PT SC M AN U

27.

TE D

26.

EP

25.

ACCEPTED MANUSCRIPT

Bender AM, Jørgensen T, Helbech B, et al. Socioeconomic position and participation in baseline and follow-up visits: the Inter99 study. European journal of preventive cardiology 2012:2047487312472076. Moayyedi P, Feltbower R, Brown J, et al. Effect of population screening and treatment for Helicobacter pylori on dyspepsia and quality of life in the community: a randomised controlled trial. Leeds HELP Study Group. Lancet 2000;355:1665-9. Hobbs FD, Delaney BC, Rowsby M, et al. Effect of Helicobacter pylori eradication therapy on dyspeptic symptoms in primary care. Fam Pract 1996;13:225-8. Bjorner J, Damsgaard M, Watt T, et al. Danish manual for SF-36, 1997.

AC C

24.

19

ACCEPTED MANUSCRIPT

Table 1 Baseline Register Data on Responders vs. Nonparticipants Individuals Invited Responders Nonresponde n=12,530 rs n=7,481 6,781 3,223 5,749 4,258 6,392 (51.0) 3,675 (49.1) 2,154 (17.2) 1,350 (18.1) 2,597 (20.7) 1,753 (23.4) 3,034 (24.2) 1,723 (23.0) 2,394 (19.1) 1,318 (17.6) 2,351 (18.8) 1,330 (17.8) 114 (0.8) 56 (0.9)

Crude OR

OR of participation 95% CI Adjuste d OR

95% CI

AC C

EP

TE D

M AN U

SC

RI PT

Control Group ref 0.64 0.61−0.68 0.63 0.59−0.67 Screened Group 1.08 1.02−1.14 1.21 1.14−1.29 Female n (%) Age n (%) 40−44 ref 45−49 0.93 0.85−1.02 0.91 0.83−1.00 1.10 1.01−1.21 51−54 1.09 1.00−1.20 1.14 1.03−1.25 1.22 1.11−1.35 55−59 1.11 1.01−1.22 1.52 1.36−1.69 60−65 1.47 1.04−2.00 Previous PUD n (%) 1.22 0.88−1.68 CCI score n (%) Low (0) 11,510 (91.9) 6,783 (90.7) ref Medium (1) 582 (4.6) 383 (5.1) 0.90 0.78−1.02 0.99 0.86−1.13 0.82 0.71−0.95 High (≥2) 438 (3.5) 315 (4.2) 0.90 0.77−1.05 Socio n (%) Employed 8,934 (71.3) 4,690 (62.7) ref Unemployed 354 (2.8) 206 (2.8) 0.90 0.76−1.07 0.94 0.78−1.12 0.66 0.62−0.70 0.62 0.58−0.67 Retired 3,242 (25.9) 2,582 (34.5) Unknown − 3 (0.0) − Education Basic school 4,063 (32.4) 2,995 (40.0) ref 1.44 1.35−1.53 1.38 1.29−1.48 Sec./short tertiary 5,875 (46.9) 3,010 (40.2) 1.38 1.27−1.49 1.26 1.16−1.38 Med./long tertiary 2,356 (18.8) 1,263 (16.9) 0.82 0.67−0.99 0.90 0.74−1.09 Unknown 236 (1.9) 213 (2.9) 1.68 1.56 1.80 1.51 1.41−1.63 Spouse yes n (%) 10,468 (83.5) 5,623 (75.2) Data above are on the period of four years prior to baseline PUD=peptic ulcer disease CCI=Charlson Comorbidity Index: CCI=0 is no comorbidity, CCI=1 is single comorbidity, and CCI ≥2 is multiple comorbidity. PUD is not included in the calculated CCI score

ACCEPTED MANUSCRIPT Table 2 Characteristics of questionnaire responders at baseline and at the 13-year follow-up Baseline data of enrolled Screened Control Group Group

13-year follow-up data (CC) Screened Group

Control Group

5,749

6,781

4,144 (72.1)

4,514 (66.6)

Male n (%)

2,845 (49.5)

3,293 (48.6)

1,986 (47.9)

2,147 (47.6)

52.2 (6.8)

52.2 (7.0)

46−58

46−58

1,424 (24.8)

1,423 (21.0)

32 (0.6)

32 (0.5)

1,584 (27.6)

1,604 (23.7)

24 (0.4)

13 (0.2)

18 (0.4)

14 (0.3)

2,100 (36.6)

2,608 (38.8)

704 (7.0)

729 (7.3)

Mean (s.d.) 25−75 percentile

Dyspepsia (aps ≥2)

n (%) Missing

Reflux (rs ≥2)

n (%) Missing

65.4 (6.6)

65.5 (6.7)

60−70

60−71

782 (18.9)

842 (18.7)

18 (0.4)

20 (0.4)

956 (23.1)

981 (21.7)

SC

Age

RI PT

Number (%)

Smoking

n (%)

24 (0.6)

19 (0.4)

Alkohol

0−5 unit/wk

2,855 (49.7)

3,332 (49.1)

1,968 (47.9)

2,130 (47.8)

5−21 unit/wk

2,617 (45.5)

3,002 (44.3)

1,976 (48.1)

2,171 (48.7)

>21 unit/wk

231 (4.0)

348 (5.1)

162 (4.0)

159 (3.6)

Missing

46 (0.8)

102 (1.5)

45 (1.1)

58 (1.2)

No use

4,496 (78.2)

5,415 (79.9)

3,022 (72.9)

3,209 (71.1)

Use < daily

993 (17.3)

1,052 (15.5)

829 (20.0)

980 (21.7)

Use daily

199 (3.5)

202 (3.02)

174 (4.2)

177 (3.9)

63 (1.1)

115 (1.7)

119 (2.9)

148 (3.3)

325 (5.7)

417 (6.2)

720 (17.4)

786 (17.4)

29 (0.5)

81 (1.2)

144 (3.5)

171 (3.8)

5,289 (92.0)

6,240 (92.0)

3,533 (85.3)

3,863 (85.6)

Use < daily

252 (4.4)

308 (4.5)

179 (4.3)

183 (4.1)

Use daily

178 (3.1)

183 (2.7)

393 (9.5)

408 (9.0)

Missing

29 (0.5)

48 (0.7)

39 (0.9)

60 (1.3)

No use

4,127 (72.6)

5,123 (75.5)

3,256 (78.6)

3,602 (79.8)

Use < daily

1,409 (24.5)

1,455 (21.5)

650 (15.7)

675 (15.0)

Use daily

133 (2.3)

135 (2.0)

145 (3.5)

136 (3.0)

Missing

34 (0.6)

69 (1.0)

93 (2.2)

101 (2.2)

Any GP visits (dyspepsia)* n (%)

207 (3.6)

216 (3.2)

119 (3.1)

121 (2.9)

Any sick leave days (dyspepsia)* n (%)

144 (2.5)

136 (2.0)

69 (1.7)

67 (1.5)

5,473 (95.2) 50.8 (8.9) 48 (54) 57 5,473 (95.2) 54.0 (8.8)

6,310 (93.1) 51.2 (8.9) 48 (54) 57 6,310 (93.1) 53.4 (9.2)

3,734 (90.1) 48.7 (9.6) 44 (52) 56 3,734 (90.1) 54.7 (8.8)

4,016 (90.1) 48.9 (9.2) 45 (52) 56 4,016 (90.1) 54.9 (8.7)

Missing Low dose ASA n (%)

Use daily Missing No use

AC C

EP

Ulcer drugs n (%)*

TE D

ASA/NSAID n (%)*

17 (0.3)

Antacids n (%)*

SF-36 Summary PCS n (%) Mean (s.d.) 25 (50) 75 percentile SF-36 Summary MCS n (%) Mean (s.d.)

54 (0.8)

M AN U

Missing

ACCEPTED MANUSCRIPT 25 (50) 75 percentile

51 (57) 60

50 (57) 59

52 (58) 60

52 (58) 60

AC C

EP

TE D

M AN U

SC

RI PT

CC=complete cases, questionnaire data from all follow-up points. Mean time of follow-up is 13.4 years (s.d.0.4, range 12.4−14.7) PPI=Proton pump inhibitor. NSAID=Non-Steroidal Anti-Inflammatory Drugs. ASA=Acetylsalicylic Acid. Low dose ASA (≤150 mg), high dose ASA (≥500 mg). High dose ASA is reported together with NSAID use. SF-36 Summary scores: PCS=Physical Component Summary and MCS=Mental Component Summary *Within the previous month

ACCEPTED MANUSCRIPT Table 3 Cumulative post-randomization risk of events for participants in HEP-FYN, registerdata for the 13-year follow-up period

Mean DDD/user S.d. H2-antagonists n (%) Mean DDD/user S.d. ASA/NSAID use n (%) Mean DDD/user S.d. Low dose ASA n (%)

Persons n (%) 13

12.4 (2.1) 181 (2.7) 134 (2.0) 59 (0.9) 322 (4.7)

0.88 (0.71−1.10) 0.86 (0.67−1.12) 0.90 (0.61−1.32) 1.24 (1.07−1.44)

0.93 (0.84−1.03)

546 (9.5)

692 (10.2)

0.93 (0.84−1.04)

22 (0.2) 15 (0.1) 30 (0.2) 4,010 (40.1) 639 1,252 1,102 (11.0) 182 605 7,608 (76.0) 778 1,404 2,240 (22.4) 1,666 1,656 221 (2.2)

28 (0.3) 11 (0.1) 24 (0.2) 4,032 (40.3) 608 1,268 1,106 (11.1) 196 531 7,629 (76.3) 796 1,421 2,402 (24.0) 1,676 1,652 301 (3.0)

0.79 (0.45−1.37) 1.36 (0.63−2.97) 1.14 (0.73−2.00)

C-urea breath test n (%)

RI PT

12.5 (1.9) 135 (2.3) 98 (1.7) 45 (0.8) 339 (5.9)

9 (0.2) 6 (0.1) 14 (0.2) 2,486 (43.2) 661 1,285

18 (0.3) 6 (0.1) 14 (0.2) 2,776 (40.9) 588 1,190

0.59 (0.27−1.31) 1.18 (0.38−3.66) 1.18 (0.56−2.00)

716 (12.5)

764 (11.3)

1.11 (1.00−1.22)

0.73 (0.62−0.87)

172 533 4,529 (78.8) 788 1,424 1,347 (23.4) 1,710 1,690 119 (2.1)

189 519 5,218 (77.0) 794 1,398 1,640 (24.2) 1,712 1,624 204 (3.0)

0.99 (0.96−1.03)

1.00 (0.92−1.08)

1.00 (0.98−1.01)

0.93 (0.89−0.98)

1.06 (1.01−1.10)

1.02 (1.00−1.04)

0.97 (0.91−1.03)

0.69 (0.55−0.86)

192 (1.9)

205 (2.0)

0.94 (0.77−1.14)

88 (1.5)

119 (1.8)

0.87 (0.66−1.15)

120 (1.2)

132 (1.3)

0.91 (0.71−1.17)

66 (1.2)

86 (1.3)

0.91 (0.66−1.25)

3,040 1,679 (16.8)

3,023 1,703 (17.0)

1.01 (0.96−1.05)

1,774 1,029 (17.9)

2,031 1,178 (17.4)

1.03 (0.98−1.09)

45 (0.4)

374 (3.7)

0.12 (0.09−0.16)

26 (0.5)

363 (5.3)

0.08 (0.06−0.13)

AC C

Mean DDD/user S.d. Hp eradication n (%) Hospital contacts persons n (%) PUD Admissions PUD Out-patient visits Upper endoscopy Total n

0.93 (0.79−1.09) 0.84 (0.69−1.02) 1.16 (0.88−1.53) 1.04 (0.92−1.17)

SC

PPII n (%)

12.3 (2.3) 295 (2.9) 216 (2.2) 97 (1.0) 474 (4.7) 1,242 (12.4)

M AN U

Cancer n (%) Oesophageal Cardia Gastric

12.3 (2.2) 274 (2.7) 182 (1.8) 113 (1.1) 491 (4.9) 1,185 (11.8)

TE D

Death n (%)

Participants enrolled (PP) Screened Control Group Group RR (95%CI) n=5,749 n=6,781

EP

Time of follow-up Mean (s.d.) PUD n (%) Uncomplicated Complicated Esophagitis n (%)

Randomized and invited (ITT) Screened Control Group Group RR (95%CI) n=10,007 n=10,004

0.98 (0.92−1.04)

1.03 (0.96−1.00)

Time of follow-up was estimated as time from index date to indexdate+13years or indexdate to deathdate CI=Confidence Interval PPI=Proton Pump Inhibitor NSAID=Non-Steroidal Anti-Inflammatory Drugs. ASA=Acetylsalicylic Acid. Low dose ASA (≤150 mg), high dose ASA (≥500 mg). High dose ASA is reported together with NSAID use DDD=Defined Daily Dose

ACCEPTED MANUSCRIPT 20,011 Randomized

10,007

10,004

Invited

Nonresponders

Nonresponders

n=4,258 639 (15.0%) died during 13-year follow-up

n= 3,223 550 (17.1%) died during 13-year follow-up

Questionnaire responders Enrolled 12,530

6,781

RI PT

5,749

Dead 42 Moved 30 No reply 487

Dead 20 Moved 19 No reply 351

No reply 298

4,821 (84%)

5-year follow–up 10,433 (83%)

13-year follow -up 8,658 (69%)

TE D

Dead 320 Moved 49 No reply 308

SC

Dead 148 Moved 92

1-year follow–up 11,581 (92%)

4,144 (72%)

6,222 (92%)

M AN U

5,359 (93%)

Hp Screening Group

Dead 183 Moved 93 Noreply 334

5,612 (83%)

Dead 390 Moved 73 No reply 635

4,514 (67%)

Control Group

AC C

EP

Figure 1 Study Population Flow – HEP-FYN A total 356 (2.8%) moved outside the Region of Southern Denmark and 1,103 individuals died before either point of follow-up.

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

TE D



Figure 2 SF-36 dimension scores and summary scores at the 13-year follow-up

AC C



EP

PF:Physical Functioning scale; SF:Social Functioning scale; RP: Role-Physical scale; RE: RoleEmotional scale; BP:Bodily Pain scale; MH: Mental Health scale; VT: Vitality scale; GH: General Health perceptions scale; PCS: Physical Component Summary score; MCS: Mental Component Summary score.



ACCEPTED MANUSCRIPT

Protocol

Long-term effect of screening and eradication of Helicobacter Pylori in the general population - HEP-FYN 13 -years follow-up

Aims of the study 1: To evaluate the long-term effect of screening and eradication of Helicobacter Pylori on

RI PT

the prevalence of dyspepsia. To assess the effect on dyspepsia related health care consumption and quality of life.

2: To investigate dyspepsia, PUD and Gastro-esophageal Reflux Disease (GERD) and

SC

their effect on quality of life, prognosis and health care expenditure.

Background

Dyspepsia is a common condition, it is reported by up to 40% of the population (1).

M AN U

Dyspepsia related health care costs are high due to consultations, examinations, and drug consumption.

The consumption of ulcer drugs has been increasing during the last decades. In 2011 9% of the Danish population had been prescribed ulcer drugs by their GP (2). There are no recent population studies describing the health related expenses due to

TE D

dyspepsia or reflux.

It is widely accepted that infection with the bacteria Helicobacter Pylori (Hp) is crucial for the progression of peptic ulcer provided that the patient has not consumed acetyl salicylic acid (ASA) or non-steroidal anti-inflammatory drugs (NSAID). Eradication of Hp prevents

EP

relapse of peptic ulcer. Furthermore it has been suggested that Hp is a risk factor for development of gastric cancer. Randomized trials from areas with a high risk of gastric

AC C

cancer indicate that Hp eradication reduces the incidence of gastric cancer (3). Functional dyspepsia is a condition that occurs 4-5 times more frequent than peptic ulcer and is often treated with ulcer medication. The significance of Hp infection for those with functional dyspepsia has proven to be limited (4). Focus has been on whether Hp infection protect against GERD and thus against esophageal cancer. Computer simulations have suggested that Hp screening and eradication at population level reduces the expenses related to upper gastrointestinal symptoms in the population, within a number of years.

Three randomized studies have been carried out (5-8) to look into this question. In two studies from the UK population groups were invited to Hp test. Individuals who were Hp

ACCEPTED MANUSCRIPT

positive (25%) were randomized to eradication therapy or placebo. In the Leeds study (5) the population follow-up is 10 years, but only 40% participated in the follow-up. This study found a reduction in dyspepsia-related costs of $117 per person over 10 years for those who were Hp positive. The Bristol study (7-8) found a significant decrease in consultations but not in dyspepsia related drug consumption in the Hp eradication group at 7-yr followup.

RI PT

This current Danish study (9) randomized at population level into intervention (n=10,007) and the control group (n=10,004) prior to invitation to the baseline visit. The strength of this study is that it gives insight into the effect of screening on population scale including the effect of information about Hp status. Current follow-up is 5 years. The Hp prevalence was

SC

17%. Hp screening caused a modest reduction in the rate of dyspepsia, consultation rate and sick leave days for dyspepsia, but no reduction in drug consumption. However, a significant reduction in incidence of peptic ulcer and ulcer complications was

M AN U

seen in the screened group compared to the unscreened group. It is important to follow this effect over time in an aging population with an expected increase in ASA and NSAID consumption.

It is shown that Hp eradication can prevent peptic ulcer in NSAID-naive patients (10). Hp

TE D

screening was not cost-effective 5 years after the intervention. This result can alter, if the rate of peptic ulcer and in particular ulcer complications is reduced further over time.

Methods Study population

EP

Ulcer complications are severe and the mortality is 10 % despite improved treatments.

AC C

1998-99 20.000 individuals, age 40-65 years, identified by their civil registration number, were allocated by a computerized randomized procedure to Hp screening group and control group. All participants received a questionnaire at inclusion, 1-yr and 5-yr follow-up. Of the 20,011 invited persons at baseline 63% were enrolled and response rate at 5-yr follow-up was 84% (figure below).

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

The 13-yr follow-up

Only individuals who participated at 5-yr follow-up (not moved or ded) receive a questionnaire at 13-yr follow-up. With a few exceptions the same basic questionnaire is

Questionnaire data:

TE D

used for all contacts.

The questionnaire included questions on abdominal symptoms (GRSR: Gastrointestinal symptom rating scale), rate of symptoms, quality of life (SF-36, EQ-5D-5L), consumption

EP

of ulcer drugs including over the counter drugs, hospital admissions, dyspepsia-related sick leave days, consultations and comorbidity). Those who did not respond to the first

AC C

invitation received a reminder and a new questionnaire after approximately two weeks.

Register data:

Data from national registers were accessed for all randomized individuals at individual level via Statistics Denmark (20). All national registers can be linked by using the unique personal identification number assigned to all Danish residents at birth or upon immigration (19). This allowed us to perform intention-to-treat analysis as well as per protocol analysis. Data were obtained for two periods. The first period comprised the four years prior to index date and was used to evaluate participant vs. nonparticipant differences. The second period was from index date to end of 13-year follow-up.

ACCEPTED MANUSCRIPT

-The Danish National Patient Register (NPR) (21): Information on ulcer-related admissions, outpatient ulcer diagnosis and comorbidity, use of upper gastrointestinal endoscopy was extracted based on International Classification of Diseases (ICD) (ICD-8 and ICD-10.). Index ulcer was predefined as the first recorded PUD diagnosis verified by upper endoscopy or surgery and further subcategorized as complicated or uncomplicated.

RI PT

Patients with a diagnosis of gastro-oesophageal cancer were identified and categorised by the gastrointestinal location. Patients with a diagnosis of esophagitis were counted. The number of hospital admissions and outpatient visits with and without a related diagnosis of PUD and the number of upper endoscopies were counted.

SC

-The Danish National Prescription Registry (DNPR): Data on consumption of reimbursable ulcer drugs, ASA, NSAID and Hp eradication therapy. Drug utilization was analyzed using

M AN U

the defined daily dose (DDD).

-The Populations Education Register: Information on highest completed education classified into educational levels.

-Information on income and Socio economic status (SES) from–

TE D

Arbejdsklassifikationsmodulet(AKM) at Statistic Denmark.

Endpoints

Primary endpoints: The overall endpoint is expenditure in Danish kroner (on ulcer drugs,

ulcer).

EP

upper gastrointestinal endoscopy, GP consultations, hospital admissions related to peptic

Data on resource consumption of these services and procedures is multiplied with relevant

AC C

unit prices, from the Danish Medicines Agency and the Danish National Board of Health, to assess a health-cost endpoint. Secondary endpoints:

-

Frequency of symptoms (dyspepsia, reflux)

-

Quality of life

-

Frequency of peptic ulcers incl. complicated ulcers

-

Incidence of esophagus- and gastric cancer.

-

Drug consumption

-

Upper gastrointestinal endoscopy

ACCEPTED MANUSCRIPT

Intended Analyses The randomized study:

Register data of participants who answered the questionnaire are analyzed. Furthermore Intention-to-treat analyses of all randomized individuals for the entire follow-up period are performed. Questionnaires are scanned and validated. Analyses are performed in STATA 11.

RI PT

Intended Statistical analyses include χ2-test, t-test and Mann-Whitney u-test as well as univariate and multivariate logistic regression analyses. All tests are 2-tailed; the levels of significance was set at p<0.05. The results of these analyses are expressed as odds ratios with 95% confidence interval.

SC

We will compare demographic data to assess whether the participants in the 13-yr followup are representative for the original study population included at baseline. Modified Charlsons index were applied for both groups in order to grade co morbidity

M AN U

based on information on International Classification of Diseases (ICD8 and ICD10) provided the Regional Hospital Discharge Register (11, 12).

Feasibility of the project

TE D

The Department of Medical Gastroenterology S, Odense University Hospital has during a number of years been engaged in issues concerning management and treatment of patients with dyspepsia. Four PhD-Theses has been accomplished, one of which was performed in relations to data gathered and analyzed at baseline and 1-yr follow-up of the

EP

current study. The collaboration with Odense Pharmaco -epidemiological Database (OPED) is close and the Department has extended experience in assessment of register

AC C

data. OPED provides key information on the individual drug consumption and thus enable us compare this information with information on diagnosis, hospital admissions and outpatient endoscopic procedures. This study is authorized to analyze register data on randomized individuals, who have not participated in the study in an anonymous form. This provides an opportunity to perform intention-to treat analyses for all randomized individuals.

Expected results The study will provide information on the long-term effect of Hp screening and eradication in a population. The study will provide information about the long-term effect on incidence of peptic ulcer in an aging population that is likely to have an increased consumption of

ACCEPTED MANUSCRIPT

ASA and NSAID. Furthermore the study will generate knowledge about the long-term prognosis of dyspepsia and reflux in the population (dyspepsia and reflux). The study does not have the power to detect significant effect on the incidence of esophageal and gastric cancer. It is however important to investigate if Hp eradication has a long-term effect on these conditions.

RI PT

Research unit

Responsible Physician, PhD student Maria Bomme, Institute of Clinical Research,

University Of Southern Denmark and Department of Medical Gastroenterology S, Odense

SC

University Hospital.

Principal Supervisor: Chief Physician, Ph.D. Clinical Associate Professor Jane Møller Hansen, Department of Medical Gastroenterology S, Odense University Hospital.

M AN U

Co-supervisor: Chief Physician, Professor of Medicine, Head of Research Unit Ove B Schaffalitzky de Muckadell, Department of Medical Gastroenterology S, Odense University Hospital.

Co-supervisor: Msc econ, PhD. Associate Professor Christian Kronborg, Centre of Health Economics Research, University of Southern Denmark.

TE D

Collaborator: Chief Physician, Professor Jesper Hallas, Clinical Pharmacology. Collaborator: Chief Physician, PhD Mette Wildner Christensen, Department of Internal Medicine, Svendborg Hospital, Denmark.

EP

Dissemination of the results: It is planned to present the results at National and

AC C

International assemblies / Congresses and published in peer reviewed journals.

Ethical aspects: The study is approved by The Regional Scientific Ethical Committee for Southern Denmark and the Danish Data Protection Agency. All data are processed and stored in accordance with the regulations of the Danish Data Protection Agency.

ACCEPTED MANUSCRIPT References 1. El-Serag H, Talley N. Systematic review: the prevalence and clinical course of functional dyspepsia. Aliment Pharmacol Ther 2004; 19:643-654.

RI PT

2. http://www.medstat.dk/en (updated 10.05.2012, accessed 31.05.2012). 3. Fuccio L, Zagari R, Eusebi L et al. Meta-analysis: Can Helicobacter pylori Eradication Treatment Reduce the Risk for Gastric Cancer? Annals of Internal Medicine 2009; 151:121-128.

SC

4. Moayyedi P, Soo S, Deeks B et al. Eradication of Helicobacter pylori for non-

M AN U

ulcer dyspepsia. Cochrane Database of Systematic Reviews 2006; (2). 5. Ford A, Forman D, Bailey A, Axon A, Moyyedi P. A community screening program for Helicobacter pylori saves money: 10-year follow-up of a randomized controlled trial. Gastroenterol 129, 1910-1917. 2006.

6. Hansen J, Wildner-Christensen M, Schaffalitzky de Muckadell O. Effect of a

TE D

community screening for Helicobacter pylori: a 5-year follow-up study. Am J Gastroenterol 2008; 103:1106-1113.

7. Lane J, Murray L, Noble S et al. Impact of Helicobacter Pylori eradication on

EP

dyspepsia, health resource use, and quality of life in the Bristol helicobacter project: randomized controlled trial. BMJ 332, 199-204. 2006.

AC C

8. Harvey RF, Lane JA, Nair P, Egger M, Harvey I, Donovan J, Murray L. Clinical trial prolonged beneficial effect of Helicobacter pylori eradication on dyspepsia consultations- the Bristol Helicobacter Project. Aliment Pharmacol There 2010; 32: 394-400.

9. Wildner-Christensen M, Hansen J, Schaffalitzky de Muckadell O. Rates of dyspepsia one year after Helicobacter pylori screening and eradication in a Danish population. Gastroenterol 2003; 125:149-154. 10. Vergara M, Catalan M, Gispert J, Calvet X. Meta-analysis: role of Helicobacter pylori eradication in the prevention of peptic ulcer in NSAID users. Aliment

ACCEPTED MANUSCRIPT

Pharmacol Ther 21, 1411-1418. 2005.

11. Charlson ME, Pompei P, Ales KL, Mackenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chron Dis 1987; 40: 373-383.

RI PT

12. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with

ICD-9-CM administrative databases. J Clin Epidemiol1992; 45: 613-619.

SC

13. Hansen J, Wildner-Christensen M, Schaffalitzky de Muckadell O. Gastroesophageal

reflux symptoms in a Danish population: a prospective follow-up analysis of

M AN U

symptoms, quality of health, and health-care use. Am J Gastroenterol 2009; 104: 2394-

AC C

EP

TE D

2403.