Portal de Periódicos da CAPES

A Community Screening Program for Helicobacter pylori Saves Money: 10-Year Follow-up of a Randomized Controlled Trial

2005; Elsevier BV; Volume: 129; Issue: 6 Linguagem: Inglês

10.1053/j.gastro.2005.09.016

1528-0012

Alexander C. Ford, David Forman, Alastair G. Bailey, A T Axon, Paul Moayyedi,

Gastroesophageal reflux and treatments

Background & Aims: Population screening and treatment of Helicobacter pylori has been advocated as a means of reducing mortality from gastric cancer, as well as dyspepsia and dyspepsia-related resource use. Previous programs have failed to demonstrate a significant effect on mortality or resource use, but follow-up was only for 1 or 2 years. We aimed to determine the effect of screening for H pylori on dyspepsia and dyspepsia-related resource use over 10 years. Methods: H pylori–positive individuals, aged 40–49 years, enrolled in a community screening program, randomized to eradication therapy or placebo in 1994, were sent a validated dyspepsia questionnaire by mail 10 years later, and primary care records were reexamined. Consultation, referral, prescribing, and investigation data related to dyspepsia were extracted. United Kingdom costs were applied to derive total cost per person (£1 = $1.8). Results: Of 2324 original participants, 1864 (80%) were traced and contacted. Of these, 1086 (47%) responded, and 919 (40%) agreed to a review of their primary care records. There was a 10-year mean saving in total dyspepsia-related costs of $117 per person (95% confidence interval [CI] = $11–$220, P = .03) with eradication therapy. Those symptomatic at baseline showed a nonsignificant trend toward resolution of symptoms at 10 years with eradication therapy (relative risk of remaining symptomatic, 0.89; 95% CI: 0.77–1.03). Conclusions: There were significant reductions in total dyspepsia-related health care costs. The savings made were greater than the initial cost of H pylori screening and treatment. Background & Aims: Population screening and treatment of Helicobacter pylori has been advocated as a means of reducing mortality from gastric cancer, as well as dyspepsia and dyspepsia-related resource use. Previous programs have failed to demonstrate a significant effect on mortality or resource use, but follow-up was only for 1 or 2 years. We aimed to determine the effect of screening for H pylori on dyspepsia and dyspepsia-related resource use over 10 years. Methods: H pylori–positive individuals, aged 40–49 years, enrolled in a community screening program, randomized to eradication therapy or placebo in 1994, were sent a validated dyspepsia questionnaire by mail 10 years later, and primary care records were reexamined. Consultation, referral, prescribing, and investigation data related to dyspepsia were extracted. United Kingdom costs were applied to derive total cost per person (£1 = $1.8). Results: Of 2324 original participants, 1864 (80%) were traced and contacted. Of these, 1086 (47%) responded, and 919 (40%) agreed to a review of their primary care records. There was a 10-year mean saving in total dyspepsia-related costs of $117 per person (95% confidence interval [CI] = $11–$220, P = .03) with eradication therapy. Those symptomatic at baseline showed a nonsignificant trend toward resolution of symptoms at 10 years with eradication therapy (relative risk of remaining symptomatic, 0.89; 95% CI: 0.77–1.03). Conclusions: There were significant reductions in total dyspepsia-related health care costs. The savings made were greater than the initial cost of H pylori screening and treatment. Gastric adenocarcinoma is the second most common cause of cancer mortality worldwide and is responsible for approximately two thirds of a million deaths annually.1Ferlay J. Bray F. Pisani P. Parkin D.M. GLOBOCAN 2002. IARC Press, IARC CancerBase No. 5, version 2.0. Lyon2004Google Scholar A systematic review of nested case-control studies found that Helicobacter pylori (H pylori) was strongly associated with distal gastric adenocarcinoma,2Helicobacter and Cancer Collaborative GroupGastric cancer and Helicobacter pyloria combined analysis of 12 case control studies nested within prospective cohorts.Gut. 2001; 49: 347-353Crossref PubMed Scopus (878) Google Scholar and there is biologic plausibility that this infection is the main cause of the disease.3Moayyedi P. Dixon M.F. Helicobacter pylori and gastric cancer implications for screening.Gastrointest Endosc Clin N Am. 1997; 7: 47-64PubMed Google Scholar A randomized controlled trial (RCT) in a high-risk population suggested that H pylori eradication reduced the risk of developing gastric cancer4Wong B.C.-Y. Lam S.K. Wong W.M. Chen J.S. Zheng T.T. Feng R.E. Lai K.C. Hu W.H.C. Yuen S.T. Leung S.Y. Fong D.Y.T. Ho J. Ching C.K. Chen J.S. Helicobacter pylori eradication to prevent gastric cancer in a high-risk region of China a randomized controlled trial.JAMA. 2004; 291: 187-194Crossref PubMed Scopus (1279) Google Scholar in a post hoc analysis of those who did not have premalignant lesions at baseline.5Parsonnet J. Forman D. Helicobacter pylori infection and gastric cancer—for want of more outcomes.JAMA. 2004; 291: 244-245Crossref PubMed Scopus (34) Google Scholar Population H pylori screening and treatment therefore has the potential to reduce gastric cancer mortality. There are accurate, inexpensive, and noninvasive tests for H pylori,6Megraud F. How should Helicobacter pylori infection be diagnosed?.Gastroenterology. 1997; 113: S93-S98Abstract Full Text PDF PubMed Scopus (113) Google Scholar and therapy for the infection is 80% to 90% effective.7Ford A.C. Moayyedi P. How can current strategies for Helicobacter pylori eradication therapy be improved?.Can J Gastroenterol. 2003; 17: b36-b40PubMed Google Scholar Once H pylori has been successfully eradicated, it is unlikely to be acquired again8Kuipers E.J. Pena A.S. van Kamp G. Uyterlinde A.M. Pals G. Kurz-Pohlmann E. Meuwissen S.G.M. Seroconversion for Helicobacter pylori.Lancet. 1993; 342 (328–321)Abstract PubMed Scopus (217) Google Scholar; therefore, unlike other programs, screening only needs to occur on a "once in a lifetime" basis. Decision analysis models9Parsonnet J. Harris R.A. Hack H.M. Owens D.K. Modelling cost-effectiveness of Helicobacter pylori screening to prevent gastric cancer a mandate for clinical trials.Lancet. 1996; 348: 150-154Abstract Full Text Full Text PDF PubMed Scopus (301) Google Scholar, 10Sonnenberg A. Inadomi J.M. Medical decision models of Helicobacter pylori therapy to prevent gastric cancer.Aliment Pharmacol Ther. 1998; 12: S111-S121Crossref PubMed Scopus (24) Google Scholar suggest that population screening and treatment may be a cost-effective use of resources, but no country has adopted this strategy. Part of the reason for this is that only developed nations can afford population H pylori screen and treat programs, and, in most of these countries, gastric cancer is becoming rare. Under these circumstances, decision analysis models suggest that H pylori "test and treat" is more cost-effective in developing countries where gastric cancer is more prevalent.9Parsonnet J. Harris R.A. Hack H.M. Owens D.K. Modelling cost-effectiveness of Helicobacter pylori screening to prevent gastric cancer a mandate for clinical trials.Lancet. 1996; 348: 150-154Abstract Full Text Full Text PDF PubMed Scopus (301) Google Scholar, 10Sonnenberg A. Inadomi J.M. Medical decision models of Helicobacter pylori therapy to prevent gastric cancer.Aliment Pharmacol Ther. 1998; 12: S111-S121Crossref PubMed Scopus (24) Google Scholar Economic models have not assessed the impact of population H pylori screening and treatment on the health service cost of managing dyspepsia. H pylori is the main cause of peptic ulcer disease11Ford A.C. Delaney B.C. Forman D. Moayyedi P. Eradication therapy in Helicobacter pylori positive peptic ulcer disease systematic review and economic analysis.Am J Gastroenterol. 2004; 99: 1833-1855Crossref PubMed Scopus (160) Google Scholar and may also be a cause of a small proportion of functional dyspepsia.12Moayyedi P. Deeks J. Talley N.J. Delaney B.C. Forman D. An update of the Cochrane systematic review of Helicobacter pylori eradication therapy in non-ulcer dyspepsia resolving the discrepancy between systematic reviews.Am J Gastroenterol. 2003; 98: 2621-2626Crossref PubMed Scopus (177) Google Scholar Population H pylori screening and treatment may therefore reduce the health service costs of managing dyspepsia, and this may partially offset the costs of the program. This would be particularly relevant to developed countries where, although the risk of gastric cancer is low, the cost to the health service of managing dyspepsia is high. We evaluated this in an RCT of population H pylori screening and treatment, and, at 2 years, there was a reduction in dyspepsia symptoms in subjects who had been allocated to eradication therapy compared with placebo13Moayyedi P. Feltbower R. Brown J. Mason S. Mason J. Nathan J. Richards I.D.G. Dowell A.C. Axon A.T.R. Effect of population screening and treatment for Helicobacter pylori on dyspepsia and quality of life in the community a randomised controlled trial.Lancet. 2000; 355: 1665-1669Abstract Full Text Full Text PDF PubMed Scopus (160) Google Scholar but no statistically significant difference in dyspepsia cost.14Mason J. Axon A.T.R. Forman D. Duffett S. Drummond M. Crocombe W. Feltbower R. Mason S. Brown J. Moayyedi P. The cost-effectiveness of population Helicobacter pylori screening and treatment a Markov model using economic data from a randomized controlled trial.Aliment Pharmacol Ther. 2002; 16: 559-568Crossref PubMed Scopus (82) Google Scholar We hypothesized that, over time, health service dyspepsia costs would be reduced in subjects allocated to H pylori eradication therapy and report the results of the 10-year follow-up from this trial. The study was conducted in the Leeds and Bradford area of West Yorkshire in northern England and was a 10-year follow-up of individuals originally recruited into the Leeds HELP study.13Moayyedi P. Feltbower R. Brown J. Mason S. Mason J. Nathan J. Richards I.D.G. Dowell A.C. Axon A.T.R. Effect of population screening and treatment for Helicobacter pylori on dyspepsia and quality of life in the community a randomised controlled trial.Lancet. 2000; 355: 1665-1669Abstract Full Text Full Text PDF PubMed Scopus (160) Google Scholar The RCT commenced in 1994 with 2-year follow-up. In the original trial, 8407 people, aged between 40 and 49 years, were recruited via their family practitioners. Subjects were chosen at random from the lists of participating primary care centers, not on the basis of presence or absence of dyspepsia. They were screened for H pylori by carbon-13-labeled urea breath test. The 2324 individuals testing positive were randomized, by a computer-generated block allocation schedule stratified by primary care center, to eradication therapy (omeprazole 20 mg, clarithromycin 250 mg, and tinidazole 500 mg, all twice daily for 7 days) or placebo. Concealment was maintained by a central trials unit, which was not involved with subject evaluation, allocating therapy to participants. Validated dyspepsia questionnaires were completed at baseline, 6 months, and 2 years,15Moayyedi P. Duffett S. Braunholtz D. Mason S. Richards I.D.G. Dowell A.C. Axon A.T.R. The Leeds Dyspepsia Questionnaire a valid tool for measuring the presence and severity of dyspepsia.Aliment Pharmacol Ther. 1998; 12: 1257-1262Crossref PubMed Scopus (117) Google Scholar and economic data on dyspepsia-related health service costs were obtained from the subjects' primary care records at 2 years.14Mason J. Axon A.T.R. Forman D. Duffett S. Drummond M. Crocombe W. Feltbower R. Mason S. Brown J. Moayyedi P. The cost-effectiveness of population Helicobacter pylori screening and treatment a Markov model using economic data from a randomized controlled trial.Aliment Pharmacol Ther. 2002; 16: 559-568Crossref PubMed Scopus (82) Google Scholar A repeat urea breath test was performed at trial completion, and all participants were informed of their H pylori status and treatment allocation. Those assigned to placebo, and those assigned to eradication therapy whose 2-year breath test results were still positive, and who wished to receive a course of eradication therapy were told to contact either their family practitioner or the trialists. For the purposes of this study, we attempted to contact all H pylori-positive individuals originally involved. To contact those involved, the primary care centers that initially recruited the individuals were visited and their computer databases accessed to obtain a current address for all participants. Those successfully traced were sent a validated dyspepsia questionnaire by mail, and informed written consent to reexamine their primary care records was requested. Nonresponders were sent a second postal questionnaire. The initial questionnaires were sent out in August 2003, and the reexamination of participants' primary care records began in January 2004. The relevant local research ethics committees in Leeds and Bradford approved the study. An original sample size of 1200 subjects in each arm of the trial was estimated to be able to detect a difference in health service dyspepsia costs of $13.50 between the 2 groups at a 90% power and 5% significance level (all costs are provided in US dollars, although based on costs in United Kingdom pounds with a conversion rate of £1 to $1.8). Assumptions underlying this calculation were that H pylori eradication therapy would cure all cases of peptic ulcer disease (with a population incidence in H pylori-positive individuals of 10% over 2 years) but have no effect on nonulcer dyspepsia. The standard deviation (SD) of the cost of peptic ulcer disease to the health service was estimated at $900, and a dropout rate of 20% was assumed.14Mason J. Axon A.T.R. Forman D. Duffett S. Drummond M. Crocombe W. Feltbower R. Mason S. Brown J. Moayyedi P. The cost-effectiveness of population Helicobacter pylori screening and treatment a Markov model using economic data from a randomized controlled trial.Aliment Pharmacol Ther. 2002; 16: 559-568Crossref PubMed Scopus (82) Google Scholar After 2 years of follow-up, the original trial did not detect a statistically significant decrease in dyspepsia-related resource use in the eradication group compared to the placebo group. Two investigators, blinded to original treatment allocation, obtained resource data. Dyspepsia-related consultations, prescribing, investigations, and secondary care referrals and admissions between 2 and 10 years were recorded, using predesigned forms, for those individuals who gave consent to reexamination of their primary care records. These data were then combined with those previously extracted at 2-year follow-up to provide complete 10-year resource use. Subsequent prescriptions for H pylori eradication therapy between years 2 and 10 were obtained from family practitioner databases and recorded for both arms of the trial. Cost data were derived from the total dyspepsia-related resource use per individual at 10 years. These were subcategorized into primary care costs (family practitioner consultations), secondary care costs (outpatient consultations, accident and emergency attendances, and inpatient admissions as a consequence of dyspepsia), costs of prescribed drugs for dyspepsia (using total defined daily doses [DDD] of acid suppression drugs and number of courses of eradication therapy), and cost of relevant investigations (barium meals, upper gastrointestinal [GI] endoscopy, and breath tests). We applied British National Formulary and United Kingdom 2002 national reference costs to resource data to obtain a total cost per person.16NHS reference costs 2002. Department of Health. Available at: http://www.dh.gov.uk/Home/fs/en. Accessed January 31, 2004.Google Scholar, 17The British National Formulary number 46, September 2003. Available at: http://www.bnf.org/bnf. Accessed April 12, 2005.Google Scholar These are detailed in Table 1. Cost data were also dichotomized according to whether the individual incurred no or any dyspepsia-related costs during 10-year follow-up.Table 1Costs Used in Obtaining Total Dyspepsia-Related Resource Use per IndividualCost (US $)Family practitioner visit49Accident and emergency attendance99Outpatient visit149Inpatient day461Proton pump inhibitor (defined daily dose)1.50H2 receptor antagonist (DDD)0.50Prokinetic (DDD)0.38Antacid (DDD)0.97Eradication therapy (1 week course)66Carbon-urea breath test40Upper GI endoscopy317Barium meal236 Open table in a new tab The presence of dyspepsia at 10 years was evaluated using the short-form Leeds dyspepsia questionnaire (S-FLDQ), which is a modified version of the 15-item Leeds dyspepsia questionnaire used in the original study.15Moayyedi P. Duffett S. Braunholtz D. Mason S. Richards I.D.G. Dowell A.C. Axon A.T.R. The Leeds Dyspepsia Questionnaire a valid tool for measuring the presence and severity of dyspepsia.Aliment Pharmacol Ther. 1998; 12: 1257-1262Crossref PubMed Scopus (117) Google Scholar The S-FLDQ assesses both the frequency and severity of 4 symptoms: indigestion, heartburn, regurgitation, and nausea (all of which are rated on a 5-point Likert scale from 0 to 4, with a higher score indicating more frequent or severe symptoms), as well as asking the subject to identify the most troublesome upper GI symptom. This questionnaire was previously validated in hospital and primary care and proved to be an acceptable, reliable, and responsive instrument for measuring both the presence and the severity of dyspepsia.18Fraser A. Qume M. Ford A.C. Redman V. Moayyedi P. Logan R.P.H. Duffy J. Wilson S. Delaney B.C. Validation of the short-form Leeds dyspepsia questionnaire (SF-LDQ) in a community setting.Gastroenterology. 2003; 124: A224Abstract Full Text PDF Google Scholar In addition, its brevity and capacity for self-administration make it convenient for use in a postal survey. Dyspepsia symptom status at 10 years was dichotomized into symptomatic or asymptomatic using the summed symptom frequency score from the S-FLDQ. A score of 4 or more (of a possible total of 16) has been shown to predict dyspepsia as diagnosed by a clinician.18Fraser A. Qume M. Ford A.C. Redman V. Moayyedi P. Logan R.P.H. Duffy J. Wilson S. Delaney B.C. Validation of the short-form Leeds dyspepsia questionnaire (SF-LDQ) in a community setting.Gastroenterology. 2003; 124: A224Abstract Full Text PDF Google Scholar Demographic data, including lifestyle factors, and socioeconomic status (as defined by occupation), baseline and 2-year symptom data, and dyspepsia-related health care costs at 2 years were stored on file from the original trial, enabling us to assess whether those individuals successfully followed up at 10 years were similar with respect to these characteristics to the original trial population. Our primary outcome measure was total dyspepsia-related resource use in those assigned to eradication therapy compared with placebo during 10-year follow-up. Secondary outcome measures included incurrence of any dyspepsia-related cost at 10 years, presence of dyspepsia at 10 years, resolution of dyspepsia at 10 years in those who were symptomatic at original trial entry, and subsequent development of gastric cancer, peptic ulcer disease, and gastroesophageal reflux disease (GERD) during 10-year follow-up, according to original treatment allocation. Differences in total and subcategorized dyspepsia-related resource use between those assigned to eradication therapy and placebo were examined using an independent sample t test and expressed as a mean difference (in US dollars) with a 95% CI. The cost data were highly skewed, whereas an independent sample t test assumes that the sample mean has a normal distribution. However, because the sample size was large, the central limit theorem suggests that an independent t test can be applied to these skewed data. This assumption was tested using bootstrap sampling, and almost identical results were obtained (data not shown). The impact of eradication therapy on incurrence of any dyspepsia-related cost, presence of dyspepsia, and subsequent development of relevant upper GI pathology at 10 years were expressed as relative risks (RR) along with a 95% CI. Subgroup analyses were performed according to sex and, with those individuals who requested subsequent eradication therapy from their family practitioner between years 2 and 10, on being informed of their H pylori status and treatment allocation, removed. For all these analyses, a 2-tailed P value of less than .05 was considered statistically significant. Demographic data were compared between groups using the χ2 test. Because of multiple comparisons, a 2-tailed P value of less than .01 was considered to be statistically significant. All statistical analyses were performed using StatsDirect statistical software version 2.2.5 (StatsDirect Ltd, Sale, Cheshire, United Kingdom), SPSS for Windows version 11.5 (SPSS Inc, Chicago, IL), and Stata version 8.0 (Stata Corporation, College Station, TX). Of the 2324 participants who were H pylori positive at baseline, 1864 (80%) could be traced. These individuals were all sent questionnaires, and 1086 of 2324 (47%) responded, 876 to the first questionnaire (Figure 1). To assess whether the individuals who agreed to participate in the follow-up study were representative of the original trial population, we compared their baseline characteristics, including sex, lifestyle factors, socioeconomic status, original treatment allocation, and presence of dyspepsia at baseline and at 2 years, as well as their dyspepsia-related health care costs at 2 years, with those individuals who were not traced or who did not respond (Table 2). Smokers and individuals with dyspepsia at baseline were less likely to be successfully followed up, whereas the opposite applied to those of a higher socioeconomic status, and alcohol users. Of those who responded, 919 of 2324 (40%) individuals gave consent to reexamination of their primary care records. Five individuals' records were missing at the time the participating primary care centers were visited, and, therefore, 914 sets of records were reviewed. There were no differences in baseline characteristics between those who gave consent to a review of their primary care records and those who refused, except that men were more likely to give consent than women (Table 3).Table 2Demographics of Individuals Who Were Successfully Contacted Compared With Those Who Were Not Contacted or Did Not RespondContacted (n = 1086)Not contacted (n = 1238)Sex (%) Male509 (47)636 (51) Female577 (53)602 (49)Lifestyle factors (%) Smoker288 (26.5)464 (37.5)aP < .01. Alcohol abstainer110 (10)195 (16)aP < .01.Socioeconomic status (%)bHousewives and members of the armed forces omitted. High340 (31)324 (26)aP < .01. Medium482 (44)573 (46) Low156 (14)186 (15)Treatment allocation (%) Eradication therapy556 (51)605 (49) Placebo530 (49)633 (51)Dyspepsia at baseline (%)439 (40)571 (46)aP < .01.Dyspepsia at 2 years (%)270 (25)270 (22)Mean dyspepsia-related cost at 2 years in US dollars (SD)102 (380)122 (460)Mean age in years (SD)55.6 (2.8)55.4 (2.9)a P < .01.b Housewives and members of the armed forces omitted. Open table in a new tab Table 3Demographics of Individuals Who Agreed to Re-examination of Primary Care Records Compared With Those RefusingAgreed (n = 919)Refused (n = 167)Sex (%) Male446 (48.5)63 (38)aP < .01. Female473 (51.5)104 (62)Lifestyle factors (%) Smoker241 (26)47 (28) Alcohol abstainer90 (10)20 (12)Socioeconomic status (%)bHousewives and members of the armed forces omitted. High294 (32)46 (27.5) Medium406 (44)76 (45.5) Low127 (14)29 (17)Treatment allocation (%) Eradication therapy476 (52)80 (48) Placebo443 (48)87 (52)Dyspepsia at baseline (%)382 (42)57 (34)Mean age in years (SD)55.5 (2.8)55.9 (2.8)a P < .01.b Housewives and members of the armed forces omitted. Open table in a new tab Data were available for 914 of the original 2324 individuals (39%), 474 of whom were assigned to eradication therapy and 440 to placebo. Baseline characteristics between the 2 groups were comparable (Table 4). There was a significant reduction in dyspepsia-related costs with eradication therapy vs placebo at 10 years, with a mean cost of $303 per person vs $420 per person, respectively. This gave a mean 10-year cost saving with eradication therapy of $117 per person (95% CI: $11–$220, P = .03), which compares with a previously reported cost difference at 2 years with eradication therapy of $21 per subject (95% CI: −$13–$54, P = .23).14Mason J. Axon A.T.R. Forman D. Duffett S. Drummond M. Crocombe W. Feltbower R. Mason S. Brown J. Moayyedi P. The cost-effectiveness of population Helicobacter pylori screening and treatment a Markov model using economic data from a randomized controlled trial.Aliment Pharmacol Ther. 2002; 16: 559-568Crossref PubMed Scopus (82) Google Scholar Between 2- and 10-year follow-up, the mean cost saving with eradication therapy compared with placebo was $74 per person (95% CI: −$11–$158, P = .087). When costs were subcategorized, it became apparent that the majority of the 10-year cost savings were a result of a reduction in dyspepsia-related prescribing in the eradication arm of the trial (mean difference in dyspepsia-related prescribing, $81; 95% CI: $19–$141, P = .01) (Table 5). When costs were dichotomized, there were significantly fewer individuals in the eradication therapy group who incurred any dyspepsia-related cost (48.5%) compared with those who received placebo (72.5%) (RR of incurring any dyspepsia-related cost, 0.67; 95% CI: 0.60–0.75, P < .001).Table 4Demographics of Individuals Who Agreed to Re-examination of Primary Care Records According to Treatment AllocationEradication therapy (n = 474)Placebo (n = 440)Sex (%) Male242 (51)201 (46) Female232 (49)239 (54)Lifestyle factors (%) Smoker125 (26)116 (26) Alcohol abstainer49 (10)41 (9)Socioeconomic status (%)aHousewives and members of the armed forces omitted. High148 (31)145 (33) Medium211 (44.5)192 (44) Low61 (13)65 (15)Dyspepsia at baseline (%)194 (41)186 (42)Mean age in years (SD)55.6 (2.8)55.5 (2.8)a Housewives and members of the armed forces omitted. Open table in a new tab Table 5Mean Dyspepsia-Related Costs and Difference in Dyspepsia-Related Costs per Individual Between Placebo and Eradication Therapy Arms by Subcategory and Total CostMean cost in eradication therapy arm (n = 474) (SD)Mean cost in placebo arm (n = 440) (SD)Mean cost difference (95% confidence interval)P valuePrimary care40 (81)47 (90)7 (−4–18).23Prescribing117 (368)198 (556)81 (19–141).01Investigations76 (176)92 (212)16 (−10–40).24Secondary care70 (313)83 (466)13 (−14–65).61Total303 (720)420 (887)117 (11–220).03NOTE. Costs in US dollars. Open table in a new tab NOTE. Costs in US dollars. Subgroup analysis by sex revealed a nonsignificant trend toward greater reductions in dyspepsia-related costs in men allocated to eradication therapy than women (mean difference in total dyspepsia-related costs, $151; 95% CI: −$12–$313, P = .07 and $88; 95% CI: −$47–$222, P = .20, respectively) when compared with their placebo counterparts. There were 54 (12%) subjects in the placebo arm and 10 (2%) subjects in the eradication therapy arm who received H pylori eradication therapy from their family practitioner between years 2 and 10. Excluding these subjects exaggerated the difference in health service dyspepsia-related costs between the eradication and placebo groups (mean difference in dyspepsia-related costs, $160; 95% CI: $48–$272). Of the 1086 individuals who returned the postal questionnaire, 1070 of 2324 (46%) gave fully analyzable symptom data at 10 years. There were no significant differences in baseline characteristics between the responders when analyzed according to original treatment allocation. The overall prevalence of dyspepsia at 10 years was 40%, and, of those with dyspepsia at baseline, 60% still had dyspepsia at 10-year follow-up. When all 1070 with analyzable symptom data were considered, irrespective of baseline symptom status, there was a trend toward a reduction in risk of being symptomatic at 10 years in those assigned to eradication therapy (221 of 547 individuals symptomatic [40%]) compared with those assigned to placebo (228 of 523 [43.5%]), giving a 3.5% absolute risk reduction (ARR), but this did not achieve statistical significance (RR of being symptomatic at 10 years, 0.93; 95% CI: 0.89–1.07). When only those individuals who were symptomatic at original trial entry were considered, there was a nonsignificant trend toward a reduction in risk of remaining symptomatic at 10 years with eradication therapy (132 of 225 individuals remained symptomatic [59%]) compared with placebo (138 of 210 [66%]), giving a 7% ARR (RR of remaining symptomatic at 10 years, 0.89; 95% CI: 0.77–1.03). When a subgroup analysis according to sex was performed, the findings were similar, with nonsignificant trends toward both a reduction in risk of being symptomatic and risk of remaining symptomatic at 10 years with eradication therapy, which were of similar magnitude for both men and women. Subgroup analysis excluding the 64 subjects who requested H pylori eradication from their family practitioner between years 2 and 10 increased the impact of H pylori eradication therapy on dyspepsia symptoms, which was of marginal statistical significance (RR of being symptomatic at 10 years, 0.76; 95% CI: 0.58–1.0, P = .05). Of the 914 individuals whose records could be accessed, identical numbers underwent upper GI endoscopy during 10-year follow-up (75 of those assigned to placebo underwent a total of 92 endoscopies, compared with 75 of those allocated to eradication therapy undergoing 90 endoscopies). None were found to have an upper GI malignancy during the 10-year follow-up period. There were a total of 12 peptic ulcers in these individuals: 5 in the eradication therapy group and 7 in the placebo arm. Of those assigned to eradication therapy, 4 individuals had received nonsteroidal anti-inflammatory drugs (NSAIDs) or aspirin, and 2 of these were also H pylori positive at the end of 2-year follow-up. Among the placebo arm, 3 individuals had received NSAIDs or aspirin, 2 were still H pylori positive, and 2 were both H pylori positive and had received NSAIDs or aspirin. An endoscopic diagnosis of GERD was made in 42 people (25 eradication therapy, 17 placebo). Neither of these differences achieved statistical significance. This is the only population-based RCT of H pylori screening and treatment to assess economic outcomes with 10 years of follow-up. We have observed a statistically significant reduction in health service dyspepsia-related costs in H pylori-positive individuals assigned to eradication therapy compared with placebo. If the prevalence of H pylori infection were 28% (or more), as in the original trial, and H pylori serology were used in place of carbon-labeled urea breath testing, these findings suggest that the savings made would be of a magnitude to cover the initial cost of instituting screening for both positive and negative individuals. This does not take into account that those testing negative may derive reassurance from screening, and their future dyspepsia-related costs may also be reduced. As we have observed, savings continue from 2 to 10 years, and it is likely that, with longer follow-up, more dyspepsia-related savings will accrue. Of note is that the sex-related difference in cost savings originally observed is no longer present.14Mason J. Axon A.T.R. Forman D. Duffett S. Drummond M. Crocombe W. Feltbower R. Mason S. Brown J. Moayyedi P. The cost-effectiveness of population Helicobacter pylori screening and treatment a Markov model using economic data from a randomized controlled trial.Aliment Pharmacol Ther. 2002; 16: 559-568Crossref PubMed Scopus (82) Google Scholar There remains a trend toward greater savings in men, but this is no longer statistically significant. There would appear to be no satisfactory biologic explanation for the earlier results, and the 10-year data would therefore seem more plausible. This study also provides valuable insight into the long-term natural history of dyspepsia, with the prevalence of symptoms both at baseline and 10-year follow-up being in the order of 40%. When those who were symptomatic at baseline were analyzed independent of treatment allocation, almost two thirds remained symptomatic at 10 years. There were absolute risk reductions in dyspepsia at 10 years in all individuals and in those who were dyspeptic at baseline when allocated to eradication therapy compared with placebo. Both these reductions are of similar magnitude to that noted in the original trial,13Moayyedi P. Feltbower R. Brown J. Mason S. Mason J. Nathan J. Richards I.D.G. Dowell A.C. Axon A.T.R. Effect of population screening and treatment for Helicobacter pylori on dyspepsia and quality of life in the community a randomised controlled trial.Lancet. 2000; 355: 1665-1669Abstract Full Text Full Text PDF PubMed Scopus (160) Google Scholar but, in this study, neither attained statistical significance, possibly because of a combination of losses to follow-up and the limited power of subgroup analyses. An alternative explanation is that there is no real difference in the prevalence of dyspepsia at 10 years between the 2 groups. No inferences can be made regarding the effect of H pylori screening and treatment on subsequent development of gastric cancer, peptic ulcer disease, and GERD for similar reasons, although these outcomes were never considered as primary concerns for this analysis. A further point to consider when analyzing the differences in symptom data between the 2 groups is that those assigned to placebo have consumed more acid-suppressant medication over 10 years than their eradication therapy counterparts, as evidenced by the magnitude of the difference in dyspepsia-related prescribing during this time. This may be a confounding factor when comparing the prevalence of dyspepsia at 10 years and could explain the nonsignificant differences observed between the 2 groups. Limitations of this study include the follow-up rate of 47%. This attrition rate is likely when following up a group of individuals randomly selected from the community over 10 years. However, the sample of participants involved in this follow-up study is broadly representative of the original trial population. Those who were successfully followed up were more likely to be nonsmokers, to be of higher socioeconomic status, and to have been asymptomatic at baseline. It is well recognized that smokers and those with chronic disease are less likely to be successfully followed up in studies such as this. These differences will reduce dyspepsia in subjects followed up and would be expected to bias the results toward the null hypothesis. In addition, the original trial participants were informed of both their H pylori status and treatment allocation at 2 years. This means that the individuals successfully followed up at 10 years were no longer masked as to treatment allocation. The possibility that the observed difference in dyspepsia-related prescribing is due to bias cannot therefore be discounted. For these reasons, the data at 2 years may be more robust. The observed differences in dyspepsia costs and symptoms between the 2 groups are increased if the individuals requesting H pylori eradication from their family practitioner are removed, so it is possible that informing the subjects of their H pylori status at 2 years may have biased the trial toward the null hypothesis. We do not have data concerning the use of over-the-counter medication among these individuals. This is unlikely to have a major impact on our findings because, in the United Kingdom, proton pump inhibitors are not currently available over the counter. H2 receptor antagonists can still be obtained either at no charge, or at greatly reduced cost, via an individual's family practitioner, compared with when purchased independently over the counter. In addition, in the original study, over-the-counter use of acid suppression medication, in the form of H2 receptor antagonists, was assessed and amounted to only $2.50 per patient per year19Moayyedi P. Mason J. Clinical and economic consequences of dyspepsia in the community.Gut. 2002; 50: 10-12Google Scholar over 2 years. Any contribution to costs is likely to be small when compared with the overall difference between the 2 groups in total health service dyspepsia-related costs. Despite these limitations, this is still a very large study, with the longest duration of follow-up, to our knowledge, to date, and the findings are important. An H pylori "search and eradicate" strategy is primarily advocated as a means of reducing the mortality from distal gastric cancer, not the community incidence of dyspepsia. Previous modeling studies comparing cost of initial screening for H pylori and treatment with projected costs of treatment of subsequent gastric cancer have suggested that this may be a cost-effective approach, depending on H pylori prevalence, probability of gastric cancer, and efficacy of eradication therapy.9Parsonnet J. Harris R.A. Hack H.M. Owens D.K. Modelling cost-effectiveness of Helicobacter pylori screening to prevent gastric cancer a mandate for clinical trials.Lancet. 1996; 348: 150-154Abstract Full Text Full Text PDF PubMed Scopus (301) Google Scholar, 10Sonnenberg A. Inadomi J.M. Medical decision models of Helicobacter pylori therapy to prevent gastric cancer.Aliment Pharmacol Ther. 1998; 12: S111-S121Crossref PubMed Scopus (24) Google Scholar These studies, however, did not take into account potential savings resulting from a reduction in dyspepsia-related costs in their analyses. Our data suggest that a "search and eradicate" strategy could be cost saving, although we have not performed a cost-effectiveness analysis. This is not the case for screening programs already implemented in many countries at the present time for other types of cancer. The authors thank the original participating primary care centers in facilitating follow-up of involved individuals and, where applicable, allowing access to primary care records: Beeston Hill Health Center, Bridge Street Surgery, Burley Park Medical Center, Burton Croft Surgery, Carlton Gardens Surgery, Chapeloak Practice, Crossland Surgery, Fountains Medical Center, Garforth Medical Center, The Grange Medical Center, High Field Surgery, Kippax Health Center, Leigh View Medical Practice, Lingwell Croft Surgery, Manor Park Surgery, Marsh Street Surgery, Meanwood Health Center, New Wortley Health Center, Park Edge Medical Practice, Robin Lane Medical Center, Silver Lane Surgery, St. Martin's Practice, The Street Lane Practice, Tinshill Lane Surgery, West Lodge Surgery, The Whitfield Practice, The Windmill Health Center, Windsor House Surgery, Woodhouse Medical Center, Woodsley Health Center, and Yeadon Health Center in Leeds; and Cullingworth Medical Center, Leylands Medical Center, Parkhurst Medical Practice, Ridge Medical Practice, and Westcliffe Medical Center in Bradford.