We have used a decision analytic model to estimate the potential clinical benefits and the associated costs of testing patients with symptoms of IBS for coeliac disease. Our results suggest that screening for coeliac disease with currently available serological tests (which have high sensitivity) can uncover nearly all cases of coeliac disease in a population of patients with symptoms of IBS, and that the resultant improvement in quality of life is achieved at a reasonable cost. Testing with TTG antibodies alone achieved most of the clinical benefit afforded by testing with a serological panel (TTG, antigliadin IgG and IgA, quantitative IgA), whose small incremental benefit was achieved at a significant incremental cost. Up-front EGD with biopsy detected a few more cases than serological testing, but at prohibitively high incremental costs per incremental case detected. Furthermore, serological testing gained more QALYs at lower costs than up-front EGD with biopsy, due to a lower incidence of complications and the lower cost of serology compared with EGD. However, in the subgroup of patients with symptoms of IBS who were already scheduled for diagnostic EGD, adding a small bowel biopsy and confirmatory antibody panel testing if histology was abnormal was an effective and cost-effective strategy for diagnosing coeliac disease. Our results can help to inform current clinical decisions, as well as point out fruitful avenues for future clinical investigation.
Our estimates of the costs per QALY gained with testing depend on various assumptions, as discussed below. Therefore, we also estimated costs per case of coeliac disease detected by testing, which do not depend on any assumptions regarding quality of life and utilities. It is not immediately apparent whether spending $4000-9000 to detect a case of coeliac disease in a 35-year-old patient with 's'-IBS is a good investment of health care dollars. It is of interest, though, that these costs per case detected are comparable with cost-effectiveness estimates in decision analytic models of uninvestigated dyspepsia, in which a minority of patients have peptic ulcer disease and a large fraction have functional dyspepsia. For instance, the cost per ulcer cured in uninvestigated dyspepsia has been estimated to be $4541 for the Helicobacter pylori test-and-treat strategy. This strategy is accepted in clinical practice. Can one compare cost per coeliac disease case detected to cost per ulcer cured? In the absence of data on patient preferences regarding this matter, we can only speculate. But we would argue that the benefits of uncovering coeliac disease in symptomatic patients are at least comparable with those of curing an ulcer in patients with dyspepsia.
We estimated QALYs based on a regression equation that predicts utilities from published SF-36 data. Direct utility measurements would certainly be superior, but they are not available at present. Nonetheless, our model suggests that even with utility gains with coeliac disease diagnosis as small as 0.005, TTG testing could be considered cost-effective at thresholds of $50 000 and $100 000 per QALY gained at coeliac disease prevalence rates in 's'-IBS as low as 2% and 1.1%, respectively. In this light, direct utility measurements of post-operative maintenance therapy in Crohn's disease are informative. An absolute reduction of 15% in the risk of Crohn's disease recurrence over 3 years, achieved by taking mesalamine, was considered by patients to yield a gain in utility of 0.005. Whether adherence to a gluten-free diet in coeliac disease is associated with an improvement in quality of life that is comparable with the improvement in quality of life represented by a 15% reduction in the risk of Crohn's disease recurrence while taking mesalamine is not known. Future research should address patient preferences regarding IBS symptoms and coeliac disease testing.
Our model's results depend on the critical assumption that uncovering coeliac disease in a patient with 's'-IBS will lead, on average, to improved quality of life. It is well-known that many people have asymptomatic or latent coeliac disease.[13,26] Adherence to a gluten-free diet is challenging and, if it did not yield benefits, adhering to a gluten-free diet would probably detract from quality of life compared with having no dietary restrictions. Because IBS is common, one might predict that at least some patients with 's'-IBS who have serological tests and small bowel biopsies suggestive of coeliac disease may have symptoms that are not attributable to coeliac disease (they have 'asymptomatic coeliac disease' as well as IBS). What is the true impact on quality of life of uncovering coeliac disease in patients with 's'-IBS? No studies have been performed in patients with 's'-IBS found to have coeliac disease to compare quality of life before and after instituting a gluten-free diet. This question must be addressed by clinical studies.
The prevalence of coeliac disease in 's'-IBS was a critical variable in our analysis. This parameter could well be different depending on the clinical population seen by individual physicians.[10,11,12] Our results suggest, however, that even at a relatively low coeliac disease prevalence in 's'-IBS, serological testing for coeliac disease is likely to be considered cost-effective. While it might be tempting to extrapolate this finding to suggest that coeliac disease screening might be worthwhile for the general population if the prevalence of coeliac disease is approximately 0.5%, we do not believe that our results should be used to advocate mass screening. In our model, it is the symptomatic improvement with coeliac disease treatment that yields gains in QALYs. This improvement is not possible in asymptomatic subjects. In order to benefit these subjects, uncovering coeliac disease must prevent long-term morbidity or mortality or decrease costs, and this benefit must be greater than any decrease in health state utility that could result from diagnosing a 'disease' that requires a major change in lifestyle. Of note, although it may seem that only patients with diarrhoea-predominant IBS symptoms may be reasonable candidates for coeliac disease testing, the data suggest a similar prevalence of coeliac disease in those with IBS-like symptoms and diarrhoea, constipation or alternating bowel habit.[10,11]
In our analysis, the antibody panel was slightly more effective than TTG testing alone, but at substantial incremental costs. These incremental costs are explained in part by the added cost of testing required for each of the 1000 's'-IBS patients ($60 more for the panel than TTG testing alone). However, the incremental cost is also affected by the fact that the panel, while more sensitive than TTG, is also less specific for coeliac disease. With the panel, therefore, more patients have false positive results and thus undergo 'unnecessary' EGD with biopsy, and are at risk for the associated complications, explaining the remaining incremental $65 spent per subject. This is an instructive illustration of the trade-off that must be assessed when one considers moving from one test to another test that is slightly more sensitive, but also less specific and more costly. However, given the significant uncertainty regarding what precisely are the differences in sensitivity and specificity between the TTG and panel tests, we would urge caution before concluding that the panel is not cost-effective compared with TTG testing alone.
In our simulation, we restricted the benefits of coeliac disease diagnosis to improvement in symptoms, and thus improvements in quality of life. Data are inconsistent regarding the impact of a gluten-free diet on the risk of small bowel lymphoma or other malignancies associated with coeliac disease.[13,23] It appears that the anaemia of coeliac disease improves with a gluten-free diet, and that problems related to infertility might be minimized with a gluten-free diet.[13,24,25] Osteopenia and osteoporosis are common with coeliac disease. Treatment with a gluten-free diet may diminish bone loss, but not reverse osteoporosis.[13,21,22] If treatment of coeliac disease prevented long-term complications, increased life expectancy, or decreased long-term costs, the benefits of coeliac disease testing in 's'-IBS would be greater than predicted in the current analysis.
Finally, it is appropriate to comment on the uncertainties inherent in decision analytic models. Often, the precise values of some inputs are not known, e.g. the true prevalence of coeliac disease in 's'-IBS and the change in quality of life after diagnosing coeliac disease. Thus, decision analyses often do not provide an irrefutable answer to a clinical question, such as 'Should all patients with IBS symptoms be tested for coeliac disease?' However, decision analyses can identify the critical variables needed for an answer, as well as identifying variables that prove not to be critical (e.g. the minimal impact of EGD cost on the cost-effectiveness of TTG vs. no testing). Thus, decision analyses can identify areas for future clinical research, and, in the absence of definitive randomized trials, they can inform policy and clinical decisions.
In conclusion, serological testing can detect most cases of coeliac disease present in a population of patients with symptoms consistent with a diagnosis of IBS at acceptable costs per case detected and per QALY gained. Up-front EGD with biopsy in all patients with symptoms of IBS is not an attractive alternative to serological testing, but it seems reasonable to add a small bowel biopsy and confirmatory serological testing in those patients with concomitant upper abdominal symptoms who are already scheduled for diagnostic EGD. Even at a relatively low prevalence of coeliac disease and modest improvements in quality of life with a gluten-free diet, serological testing is likely to be cost-effective. If coeliac disease treatment prevented long-term complications or increased life expectancy, the benefits of coeliac disease testing in suspected IBS would be even greater than we estimated. Future research should address patient preferences regarding IBS symptoms and coeliac disease testing, as well as determining whether diagnosing coeliac disease in patients with IBS symptoms indeed yields meaningful improvements in quality of life.
CLICK HERE for subscription information about this journal.
There was no financial support for this study.
Correspondence to: Dr U. Ladabaum, Division of Gastroenterology, S-357, Box 0538, University of California, San Francisco, 513 Parnassus Avenue, San Francisco, CA 94143-0538, USA. E-mail: email@example.com
Aliment Pharmacol Ther. 2004;19(11) © 2004 Blackwell Publishing
Cite this: Serological Testing for Coeliac Disease in Patients With Symptoms of Irritable Bowel Syndrome: A Cost-Effectiveness Analysis - Medscape - Jun 01, 2004.