PEDIATRICS Vol. 118 No. 2 August 2006, pp. 594-602 (doi:10.1542/peds.2005-2123)
ARTICLE |
Screening for Celiac Disease in Asymptomatic Children With Down Syndrome: Cost-effectiveness of Preventing Lymphoma
Department of Pediatrics, Indiana University School of Medicine, Indianapolis, Indiana
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
BACKGROUND. Studies demonstrate an increased prevalence of celiac disease in persons with Down syndrome, leading some organizations and authors to recommend universal screening of children with Down syndrome. However, many children with Down syndrome are asymptomatic, and the long-term implications of screening are unknown. The complication of celiac disease that leads to mortality in the general population is non-Hodgkin's lymphomas.
OBJECTIVES. The purpose of this research in asymptomatic children with Down syndrome was to (1) calculate the number needed to screen to prevent a single case of lymphoma and (2) present a cost-effectiveness study of screening.
METHODS. We constructed a decision tree using probabilities derived from the published literature for Down syndrome or from the general population where Down syndrome-specific data were not available. Celiac disease was determined by serologic screening and confirmation with intestinal biopsy. Sensitivity analysis was used to alter probability estimates affecting the cost of preventing lymphoma.
RESULTS. Using our baseline values, the no-screen strategy is dominant; that is, screening not only costs more but also results in fewer quality-adjusted life-years. A screening strategy costs more than $500000 per life-year gained. Screening all asymptomatic children with Down syndrome for celiac disease costs almost $5 million to prevent a single case of lymphoma.
CONCLUSION. These analyses do not support the cost-effectiveness of screening, and more data are needed before recommendations to screen asymptomatic children with Down syndrome for celiac disease can be made.
Key Words: Down syndrome • celiac disease • cost-effectiveness
Abbreviations: CD—celiac disease • DS—Down syndrome • SIR—standardized incidence ratio • SMR—standardized mortality ratio • Ig—immunoglobin • GFD—gluten-free diet • EMA—endomysial antibody • tTG—tissue transglutaminase antibody • CHD—congenital heart disease • CI—confidence interval • QALY—quality-adjusted life-year
Celiac disease (CD) is an immune-mediated enteropathy caused by a permanent sensitivity to gluten in genetically susceptible individuals.1 Many studies have demonstrated the increased prevalence of CD in Down syndrome (DS). It is estimated that 4.6% to 13%2–5 of children with DS have CD. Because of this increased prevalence, recommendations for universal screening of children with DS for CD have been made in the literature,2–4 by professional organizations,1 and in health care guidelines.6 In the general population, others have been more cautious in their call for screening.7–9 Maki et al,7 stated, "Further studies of the effect of asymptomatic CD, including cost-effectiveness evaluations, are needed before population-based screening studies can be recommended." Hoffenberg et al9 concluded that, "Additional studies on routine screening of at-risk groups for evidence of CD are needed before evidence-based recommendations can be developed." To this point, there have been no cost-effectiveness studies to evaluate screening individuals with DS (an at-risk population). In this study, we review the literature regarding CD and DS and then use decision analysis to look at outcomes of screening. Finally, we discuss how this evidence base can be used in informing families and patients and its implications in the formulation of health policy.
At the time of diagnosis of CD, between 44% and 69%2,3,10 of children with DS have abdominal complaints symptomatic of CD; between 11% and 39%2,10 have signs of the disease, such as growth failure or anemia; and 17% to 50%2,10,11 are asymptomatic. Although in the general population, the ratio between symptomatic and silent forms of CD is 1:8,12 the ratio is reversed in DS to 4:1, that is, more overt clinical manifestations than silent disease.2 A low threshold for testing for CD in symptomatic children, youth, and adults with DS is warranted to treat the signs and symptoms of the disease.
Sorting out the benefits of treatment and risk of complications from CD, however, is complex when considering screening for asymptomatic disease as opposed to testing when clinically indicated. There is a significant burden associated with treatment of CD involving stringent dietary restrictions (gluten-free diet) that are both difficult and costly. Furthermore, there is a substantial gap in knowledge in asymptomatic CD as to the risks of morbidities and the effect of dietary management on long-term outcomes, and there are no studies on long-term outcomes in individuals with DS. Adding to the complexity is that any potential benefits of treating silent disease may be limited by adherence to the diet which ranges from 50% to 73%13–15 in those diagnosed with CD. Furthermore, 
1 study16 suggests that children who have symptomatic CD during childhood but are not diagnosed until adulthood do not have increased mortality over those who first develop symptoms and are diagnosed as adults.
The complication of CD that leads to mortality in the general population is non-Hodgkin's lymphomas, primarily gut lymphomas. Gut lymphomas associated with CD peak during the sixth and seventh decades of life. The prevalence of gastrointestinal lymphomas in adults, however, is rare, and the risk in persons with DS may be even lower than the general population. Studies vary markedly in the risk of lymphomas in persons with CD,13,16 with a recent case-control study by Catassi et al8 showing that the CD-associated attributable risk of non-Hodgkin's lymphoma was not significantly higher than 0. Although children with DS are at increased risk of malignancies,17 2 studies in adults did not show significant differences between individuals with DS and the general population in standardized incidence ratios (SIRs)17 or standardized mortality ratio (SMR)18 for lymphoma, non-Hodgkin's lymphoma, stomach, small intestine, or colon cancers. In a third study, adults with DS were at decreased risk from malignancies.19 No data on risk of lymphoma in people with both CD and DS exist.
Decision analysis20,21 is a tool for weighing alternative courses of action in terms of their potential benefits and liabilities. In this case, the potential benefit of preventing gastrointestinal malignancy by detecting asymptomatic CD must be balanced against the cost and quality-of-life decrement associated with screening for and treating CD. Making these comparisons requires several assumptions. Decision analysis allows us to test our assumptions over a wide range of variables when precise data are unknown to inform a health policy about questions such as the cost-effectiveness of screening asymptomatic children with DS for CD.
The purpose of this study was to: (1) perform cost-effectiveness analyses of screening asymptomatic children with DS for CD to prevent non-Hodgkin's lymphoma, the major complication of CD that causes mortality; and (2) calculate the number of asymptomatic children with DS needed to screen to detect 1 case of the lymphoma.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Using standard decision modeling techniques,20,21 we constructed a decision tree using DATA 4.022 (a decision analysis software package) to evaluate the cost-effectiveness of screening for CD in asymptomatic children with DS (Fig 1). The decision tree depicts the options of screening and not screening. Whether screened or not, the child may or may not have underlying CD. After the screening branch, a child may have a positive or negative screen. If the screen is positive, the child is assumed to undergo small bowel biopsy (gold standard), which will either confirm or disprove the diagnosis of CD. If CD is confirmed, the child is assumed to be placed on a gluten-free diet (GFD). For all people with DS, there is a risk of developing lymphoma. This risk is adjusted by whether or not there is underlying CD and/or treatment with GFD.
|
Variables
Table 1 outlines the variables, the baseline values, and range of values used in the sensitivity analysis and the sources used in constructing the model. The text of the methods describes how we derived each value. In general, we reviewed the literature for the entire range of probabilities, then took the median value as our baseline, with the range defining the values for our sensitivity analysis. The probabilities were derived for DS when published literature was available or from the general population where DS-specific data were not available. In each instance where the general population data were used, conservative estimates were used (ie, values that biased toward screening). Cost data were derived from published literature, Medicare or Medicaid claims, and our own institution. We used costs where available and charges when cost data were not available. Where specific data were not available, the assumptions that were made are detailed in the text below. The analysis was conducted from a societal perspective in that costs to all parties were considered, as well as the length and quality of life for effected individuals. We did not consider secondary costs, such as time lost from work or transportation costs. Ignoring these costs tended to bias the model in favor of screening.
|
Cost of Screening
The majority of studies of screening use an immunoglobin (Ig) A antibody to endomysial (EMA), or more recently, tissue transglutaminase (tTG) antibodies. Because the costs vary considerably across laboratories and across studies, we used a baseline median value, then sensitivity analysis across the range of reported values. Although there is some evidence that tTG may be more sensitive and more specific than the EMA, other studies show comparable results (see "Sensitivity and Specificity" section below), and the costs overlap. For example, in one study23 the range of costs at 3 reference laboratories in the summer of 2001 for IgA EMA was $45 to $99, whereas IgA tTG ranged from $85 to $164. More recently, the cost of IgA EMA alone in our reference laboratory was $38 (Clarian Hospital Laboratories, personal communication, 2004), and in the Celiac Disease Study Group Service Laboratory it was $30.24 2005 Medicare reimbursement for "antibody detection, NOS [not otherwise specified]" is $22.65.25 People with CD also frequently have IgA deficiency that may result in false-negative tests, so some clinicians also order IgA levels or a combination of antibody tests when screening for CD. We did not, however, include the additional costs of blood draw or other tests that might be ordered with the screen for CD. By ignoring the additional costs that may be associated with serologic screening, we bias the model in favor of screening.
Cost of Endoscopy and Small Bowel Biopsy
Once an individual screens positive for CD, he or she will undergo endoscopy and small bowel biopsy, which is considered the gold standard for the diagnosis of CD. It is important to include these "induced costs" after initial screening, because they are incurred only if screening is done.20,21 Although these costs may also be incurred when testing symptomatic children, the financial and clinical implications would be different. The cost of endoscopy and small bowel biopsy varies greatly across sites, and many studies are from other countries with reported costs that include the conversion of foreign currency into dollars. Costs in the literature range from $300 to $1800. We used the 2005 Medicare reimbursement of $44625 and the Medicaid reimbursement in our state for the physician fee of $18126 for a total of $627 as the baseline. Given the wide range of values, we used conservative (low) estimates, which would again bias our model in favor of screening. For example, we did not include the costs of any complications or the probability of death with endoscopy, because although these costs are high, the events are rare.
Cost of a GFD
GFD is an additional induced cost of screening. The incremental costs of a GFD estimated from a survey of 25 patients resulted in a lifetime incremental cost of $44000.23 Greco et al15 estimated the costs to the Italian Health Ministry of providing a GFD to an adolescent was 1490 European currency units (which is roughly $1800) per year. We also calculated costs by searching 2 online sites, one a GFD product store27 and the other an Internet grocery store,28 for 10 commonly used food products. At each site, an average-priced, name brand product was selected with the cost figured per ounce. Whenever possible, similar sizes were compared to avoid bias related to buying in bulk. The cost per ounce of the GFD product was then divided by the cost of the regular product. The ratio of these 10 products was then averaged for comparison. On average, the gluten-free products cost 2.4 times more (with a range of 1.3–4.2) than the standard products. Using this ratio, we then calculated the incremental cost to a theoretical person who spent $50 per week on groceries for 52 weeks per year for 50 years, if 10% or 20% of their products were purchased as GFD. The incremental cost ranged from $31200 to $62400, depending on the discount rate applied. We used $44000 as a baseline estimate and conducted sensitivity analysis across the range of values.
Cost of Treating Lymphoma
The costs of lymphoma were taken from the National Cancer Institute Web site.29 The costs of lymphoma the first year after diagnosis total $17217. We did not include follow-up costs for ongoing medications or diagnostic tests, but these costs drop dramatically after the first year of treatment. We did include costs up to $1000000 in our sensitivity analysis.
Probability of CD in Asymptomatic Children With DS
Prevalence of CD in the general US population varies from 0.312% to 0.949%.30 The prevalence of CD in children with DS ranges from 4.6% to 13%.2,3,10,31,32 At the time of diagnosis, between 44% and 69%2,3,10 of children with DS have abdominal complaints symptomatic of CD; between 11% and 39%2,10 have signs of the disease, such as growth failure or anemia; and 17% to 50%2,10,11 are asymptomatic. Although in the general population, the ratio between symptomatic and silent forms of CD is 1:8,12 the ratio is reversed in DS to 4:1; that is, there are more overt clinical manifestations than silent disease.2 To derive our baseline values, we used the prevalence range from the literature of 4.6% to 13%, and we conservatively estimated that half (50%) would be asymptomatic and so would only be found by screening. This gave a range of prevalence in asymptomatic children with DS of 1% to 6.5%, with the median value of 3.3%.
Some children may have mildly symptomatic but undetected CD. From the disease detection standpoint, these children would be asymptomatic. However, early detection could potentially result in improved quality of life. Among children with DS, there is a high prevalence of gastrointestinal symptoms, and eliciting this information depends on how thorough a history is taken. We assume that children with symptoms that significantly impair quality of life will come to the clinician's attention.
Life Expectancy in DS
Median life expectancy for children with DS has been increasing over the past 20 years, primarily because of early treatment of congenital heart disease (CHD) and childhood cancers. The median life expectancy was 49 years in 1997 (in whites) and had risen at a rate of 
1 year of life expectancy per year33 in the population with DS that included individuals with CHD. Median life expectancy has not changed so dramatically for individuals without CHD, rising from a median of 49 to 52 years between 1983 and 1997.19 In a different study, the average life expectancy in adults (>40 years) with DS was 55.8 years.34 To use conservative estimates of life expectancy, we used 56 years as the baseline for individuals with DS. We examined a range with the lowest value being median life expectancy (52 years) in 1997 and extrapolate a high value (60 years) from the rate of life expectancy improvement of those without CHD.
Life Expectancy for Persons With DS Who Develop Non-Hodgkin's Lymphoma
The life expectancy for persons with DS and CD who develop non-Hodgkin's lymphoma is not known. In the general population, the 30-month survival associated with lymphoma in those without CD was 52% (95% confidence interval [CI]: 29%–71%) and poorer in those with CD where survival was only 13% (95% CI: 4%–27%).35 In another study, survival of patients in the general population with non-Hodgkin's lymphomas with CD was 50% survival at 2 years.8 The mean age of presentation for lymphoma in individuals with CD was 61 years, with a range of 36 to 82 years.35 The mean age of presentation in individuals with DS is not known. We use the range of 36 (youngest age in the previous study of presentation with lymphoma) to 56 years (life expectancy for DS from paragraph above) to assign a median age of 46 years for the development of lymphoma. We then conservatively allow a survival rate of 52% at 2.5 years. This gives a baseline life expectancy value for persons with DS who develop non-Hodgkin's lymphoma of 48.5 years with a range of 38 to 58 years.
Risk of Lymphoma
The probability (incidence) of non-Hodgkin's lymphomas in the general population varies by age and race, so it is generally reported in rates adjusted for both. White, non-Hispanic men have the highest rates of 20 per 100000 at ages 30 to 54 years and 47 per 100000 at ages 55 to 69 years, with an overall incidence rate of 20 per 100000, or 0.02%.36 The risk of non-Hodgkin's lymphomas in adults with DS may be even lower than the general population. Although children with DS are at increased risk of malignancies, 2 studies in adults did not show significant differences between individuals with DS and the general population in SIRs17 or SMR18 for lymphoma, non-Hodgkin's lymphoma, stomach, small intestine, or colon cancers. In a third study, adults with DS were at decreased risk; that is, the standardized mortality odds ratio for lymphatic and hematopoietic tissue was 0.17 (95% CI: 0.13–0.23) and for malignancies other than leukemia was 0.044 (95% CI: 0.04–0.06). In this same study, the SIR for all solid tumors for persons with DS was 0.50.19 We used a risk of lymphoma of 2% (the gender [men] and race [white] with the highest rates at age 30-54 years) as our baseline value. For the lower end of our range, we multiplied the SIR for solid tumors in DS (0.50) times our baseline value; for the upper end of our range, we used the risk of non-Hodgkin's lymphoma in the highest race and gender group (white men) at ages 55 to 69 years. This gives us a range for the risk of lymphoma of 0.01% to 0.047%.
Relative Risk of Lymphoma With CD
Several studies have looked at the additional risk of lymphoma attributable to CD, but no studies on the risk of lymphoma in adults with both CD and DS exist. A recent Evidence Report30 by the Agency for Healthcare Research and Quality summarizes the data. With the exception of 1 small study, values for the SIR varied from 2.7 to 6.3. The exception was an older, smaller study with an unusually high SIR of 42.7. We chose the largest of the SIR values (6.3) that was not an outlier as our baseline value of the relative risk of lymphoma among children with DS and CD.
Effectiveness of GFD in Preventing Lymphoma
Although people who are symptomatic for CD are at increased risk for non-Hodgkin's lymphoma, there is some evidence that following a strict GFD substantially decreases13 the risk, perhaps even to the baseline population level.16 However, a GFD is somewhat complicated, and not all individuals with CD follow a strict diet. Adherence to a strict diet ranges from 50% to 73%.13–15 We used, as baseline, a median value of 62% strict adherence and assumed that strict adherence decreased the risk to the population level. In the sensitivity analysis, we also allowed the assumption of 100% compliance.
Sensitivity and Specificity of Screening
There are basically 3 types of serologic antibody tests available for the diagnosis of CD; the older anti-gliadin antibody is less sensitive and specific than either the EMA or the more recent tTG. These tests measure antibody reactivity (usually IgA but also may measure IgG) with immunofluorescence (EMA) or enzyme-linked immunosorbent assay (tTG). They have been further improved on by using different substrates. In a recent study funded by the Agency for Healthcare Research and Quality,30 data from 18 studies of children were pooled to determine that IgA EMA (monkey embryo) has a sensitivity of 96.1 (95% CI: 94.4–97.7) and a specificity of 97.4 (95% CI: 96.3–98.2). Using a different substrate (human umbilical cord blood), the sensitivities remain about the same, but the specificity nears 100%. In children, for IgA tTG (guinea pig), the pooled estimate showed similar sensitivities (93.1; 95% CI: 88.8–95.9) and specificities (96.3; 95% CI: 93.1–98.0) to IgA EMA-ME (monkey embryo). In a small number of studies, using a human recombinant substrate, both the sensitivity and specificity of the tTG improved (sensitivity: 95.7 [95% CI: 90.3-98.1]; specificity: 99 [95% CI: 94.6-99.8]). For the purposes of this study we use these values as our baseline and range.
Utility of a GFD
Utility is the relative desirability of a health state where perfect health is assigned a utility of 1 and death a utility of 0. Mein and Ladabaum37 used short-form 36 from a Scandinavian study of adults37 to calculate the utility of treated CD. Treated CD resulted in a utility decrement of 0.009. Hence, we applied a utility for perfect health of 1.0 to persons with DS without GFD and used a utility of 0.99 as our baseline value for those placed on a GFD.
Utility of Lymphoma
No studies of utility of non-Hodgkin's lymphoma in the United States were found. In a Dutch study, utility for initial treatment, progression, or no response to treatment was 0.60, whereas utility for progression-free treatment was 0.8138 in diffuse large cell lymphoma. We, therefore, used a median value of 0.70 and performed sensitivity analysis over the range of 0.60 to 0.81.
Outcomes
The decision tree in Fig 1 represents the chance events that can follow the decision to screen or not. Each path (through the branches to the end points) represents 1 possible sequence of events. Each branch point is associated with a probability (described above and in Table 1) or a probability formula. For example, the probability of lymphoma among patients with CD who are on a GFD is the baseline risk of lymphoma multiplied by the relative risk of lymphoma among patients with CD multiplied by the efficacy of a GFD in preventing lymphoma.
At the end of each branch in the decision tree, we calculated 3 outcomes. The first was quality-adjusted life-years (QALYs) gained. QALYs are a measure of the length of life adjusted for the desirability of a health state (life expectancy times utility). The utility is the product of the utilities of the individual outcomes experienced on a particular branch (eg, utility of GFD times utility of lymphoma). Perfect health is assigned a utility of 1 and death a utility of 0. So 1 year of life in a health state with a utility of 75 would represent 0.75 QALYs. The second measure is costs. The cost term at the end of a branch is the sum of the costs incurred along the path (eg, the cost of screening plus the cost of testing plus the cost of a GFD). The expected utility and the expected cost of each strategy are determined by multiplying the probabilities along a particular path by the utility or cost term at the end of the path. Then these products were added across all of the paths following a particular strategy.
Using expected cost and QALYs, we calculated the incremental cost-effectiveness ratio. This is the dollars spent per QALY gained. It is calculated by dividing the increase in cost for the more expensive strategy by the change in QALYs with the more expensive strategy. The third outcome measure is the number of asymptomatic individuals undergoing tests (screening and subsequent small bowel biopsy) to prevent 1 case of lymphoma.
Sensitivity Analyses
Because the values used in the base case are derived from the literature and other sources, they may vary over a range of estimates. We, therefore, performed sensitivity analyses in which we varied the value of each variable across its range to determine whether our results were robust across these values. We conducted 1-way sensitivity analysis (varying 1 estimate at a time) on all of the variables and 2-way sensitivity analysis (varying 2 estimates simultaneously) on selected variables of interest. To identify thresholds at which the optimal strategy would change from "no screening" to "screening," we defined the willingness-to-pay to be $50000 per QALY.39
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
The results of the base-case analysis (using the baseline values) of the decision tree are shown in Table 2. The first column shows the strategies of screening and not screening. The "Cost" column shows the average cost of each strategy. The "Incremental Cost" is the difference in cost between the 2 strategies, that is, on average it is $1450 per child screened more expensive to screen than not to screen. The "Effectiveness" column shows the average effectiveness of each strategy expressed as QALYs expected per child screened. The "Incremental Effectiveness" is the difference in effectiveness between the no screen and screen strategies. We found that screening results in a decrease in QALYs. The "Incremental Cost/Effectiveness" is the ratio of the difference in cost and the difference in effect. Using our baseline values, the no-screen strategy is dominant; this means that screening not only costs more but also results in fewer QALYs.
|
If we do not include the quality of life adjustments, we can calculate costs per life-year saved. The results are shown in Table 3. Incremental costs are much higher for the screen strategy. Incremental effectiveness, that is, change in life expectancy, is higher, but the average increase is extremely small in the screen strategy. The "Incremental Cost/Effectiveness" shows that a screening strategy costs more than $500000 per life-year gained. By comparison, the maximum cost per QALY gained that is conventionally considered cost-effective is $5000039 or one tenth of the cost of screening for CD in asymptomatic patients with DS.
|
We can also look at the cost of preventing 1 case of lymphoma (Table 4). Incremental cost remains $1450. Effectiveness, that is, screening and treating CD, is slightly higher in the screening strategy. Specifically, the risk of lymphoma is reduced by 3 in 10000. The incremental cost/effectiveness ratio shows that screening all asymptomatic children with DS for CD costs almost $5 million to prevent a single case of lymphoma. Another way to look at these data is that 3319 children would undergo screening tests, 137 children would have biopsies, and 105 children would be placed on a GFD for every case of lymphoma prevented.
|
Sensitivity Analysis
In 1-way sensitivity analyses across all of the variables, we found that within the reported range of values, the incremental cost-effectiveness of screening never fell below our threshold of $50000 per QALY. In fact, screening reduced QALYs (ie, resulted in poorer outcomes) across all of the sensitivity analyses except 2. When the relative risk of lymphoma among children with DS and CD exceeded 14.9, screening offered greater QALYs but at a cost of $41 million per QALY, orders of magnitude higher that the usual $50000 benchmark.39 When the utility of maintaining a GFD was increased to 1.0 (ie, no disutility for the diet), screening offered higher QALYs but at a cost of $163000 per QALY (again, much higher than standard benchmarks). Two-way sensitivity analyses did not uncover conditions under which screening was favored.
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Our study found that screening asymptomatic children with DS for CD is not cost-effective in preventing lymphoma, the complication that is associated with mortality. In fact, screening in this circumstance is one of the rare strategies that not only costs more but results in worse QALYs. Even if we ignore the negative impact on quality of life associated with a GFD, screening the children who are asymptomatic is extremely expensive, many times greater than the maximum cost per QALY gained that is generally considered cost-effective. Not only will children with DS undergo unnecessary blood tests but, more importantly, procedures that hold additional risks that were not included in our study.
There are several limitations to our study. First and foremost, there are uncertainties inherent in all decision analytic models, because the precise values are not known. In this case, DS-specific data are very limited. To test these uncertainties, we performed sensitivity analyses over a wide range of values and came to the same conclusions. Second, we did not include morbidities in our analyses. If treatment of asymptomatic CD decreased other medical costs, we could be underestimating the benefits of screening. Finally, the protective effects or benefits of CD in persons with DS are unclear. For example, persons with DS are at a higher risk of mortality from ischemic heart disease.18 Yang et al19 report that CD may have a protective effect on ischemic heart disease in the general population, but we could find no reports in DS.
This study, like many studies, points to the gaps in our knowledge. Overall, DS-specific data are missing. In the general population, we have no data to say that earlier case finding leads to better outcomes, especially in childhood. Case finding in individuals with DS may be more difficult, because the onset of signs and symptoms of CD may be subtle or underreported. Children with DS may be less able to articulate their symptoms and have been reported to have increased pain thresholds (M. Bull, personal communication, 2005). Hence, children who are considered asymptomatic may in fact have unappreciated signs or symptoms. We allow, however, in our sensitivity analysis for a wide range of probability of disease, and the model remains robust. In fact, children with DS frequently do have abdominal symptoms, so even targeted testing may be quite expensive, but the trade-off with symptomatic children is that treatment may improve quality of life. Evaluating the cost-utility trade-offs in this setting is beyond the scope of the present study.
Our analysis focused on the prevention of lymphoma, because it is a clinically important outcome that has been a motivation in the recommendation to screen asymptomatic individuals for CD.2,3 Using lymphoma as the outcome, we were able to assume that the quality of life for adults with DS who developed lymphoma would be the same as that reported in the literature for adults who do not have DS. Although other conditions may be responsive to treatment, using conditions such as osteopenia, anemia, growth, and general well-being as an "outcome" is more problematic. Not only is it unknown as to how long asymptomatic individuals found by screening will take to develop subclinical findings, but it is also unclear as to when findings will become significant enough to effect quality of life or clinical outcomes. For example, the clinical significance of the osteopenia, often seen in persons with CD and also seen in persons with DS, is unclear. There is no evidence in the literature of increased risk of fractures in persons with CD40 in the general population, and we could not find a single study of fractures in DS. There have been only a few studies on quality of life in persons with CD in the general population. These studies show variable results, depending on what is measured. There are no studies of quality of life in persons with DS, and none of the existing scales include measures that may be important to individuals with DS, such as feelings of inclusion with their peers. Prospective studies weighing the costs (morbidities, family burden, and economic costs) and benefits (growth and quality of life) of screening and treating asymptomatic CD are much needed.
Our study used conservative estimates in the decision analysis, that is, bias toward performing the screening test. One example is our estimation of the prevalence of CD in asymptomatic children with DS. If the true prevalence in asymptomatic children is actually lower, then the positive predictive value of the screening tests (which is primarily influenced by the specificity and the prevalence of CD in a population) will fall dramatically.30 This means that there will be many more false positives on the screening tests resulting in many more children unnecessarily undergoing small bowel endoscopy and biopsy.
Professional organizations and advocacy groups who make recommendations for therapies or treatments may be, at times, at odds with payors, such as Medicaid. For payors, recommendations must show cost-effectiveness, especially in these times of economic constraint. In the absence of large prospective cohort or randomized, controlled trials, decision analyses such as this may be best suited for weighing complex decisions to inform health policy. Our analysis shows that, in this instance, where screening seems to be both costly and detrimental to the child, both payors and child advocates may be aligned in opposing screening of asymptomatic children with DS for CD.
This study, from a societal perspective, does not preclude a family centered approach to the care of children and families with DS. We feel that families should be given information about what is known and not known about CD in persons with DS. This allows families to weigh the costs (time, hassle, and dollars) with the benefits for their own family. It also educates families about subtle symptoms to report, which may lead to testing for CD. Physicians in the medical home should also be informed about the signs and symptoms of CD in children with DS so that testing when appropriate can be done. These views are consistent with the National Institutes of Health Consensus Development Conference Statement on Celiac Disease.41
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Screening of asymptomatic children with DS for CD to prevent lymphoma is costly and does not improve QALYs. Screening of asymptomatic children with DS for CD should not be recommended at this time. Large prospective studies in children with DS are needed to fill in the gaps of our knowledge about CD.
|
FOOTNOTES |
|---|
Accepted Mar 7, 2006.
Address correspondence to Nancy L. Swigonski, MD, MPH, Children's Health Services Research, Riley Hospital for Children, Riley Research, Room 344, 699 West Dr, Indianapolis, IN 46202-5140. E-mail: nswigons{at}iupui.edu
The authors have indicated they have no financial relationships relevant to this article to disclose.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
- Hill ID, Dirks MH, Liptak GS, et al. North American Society for Pediatric Gastroenterology HaN. Guideline for the Diagnosis and Treatment of Celiac Disease in Children: Recommendations of the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition. J Pediatr Gastroenterol Nutr. 2005;40 :1 –19.[CrossRef][Web of Science][Medline]
- Bonamico M, Mariani P, Danesi HM, et al. Prevalence and clinical picture of celiac disease in Italian down syndrome patients: a multicenter study. J Pediatr Gastroenterol Nutr. 2001;33 :139 –143[CrossRef][Web of Science][Medline]
- Book L, Hart A, Black J, Feolo M, Zone JJ, Neuhausen SL. Prevalence and clinical characteristics of celiac disease in Downs syndrome in a US study. Am J Med Genet. 2001;98 :70 –74[CrossRef][Web of Science][Medline]
- Carlsson A, Axelsson I, Borulf S, et al. Prevalence of IgA-antigliadin antibodies and IgA-antiendomysium antibodies related to celiac disease in children with Down syndrome.
Pediatrics. 1998;101
:272
–275
[Abstract/Free Full Text] - Bonamico M, Mariani P, Danesi HM, et al. Prevalence of celiac disease in Down's syndrome. J Gastroenterol Hepatol. 2001;33 :263 –267
- Cohen W. Health care guidelines for individuals with Down syndrome: 1999 revision. Down Synd Q. 1999;4 :1 –15
- Maki M, Mustalahti K, Kokkonen J, et al. Prevalence of Celiac disease among children in Finland.
N Engl J Med. 2003;348
:2517
–2524
[Abstract/Free Full Text] - Catassi C, Fabiani E, Corrao G, et al. Risk of non-Hodgkin lymphoma in celiac disease.
JAMA. 2002;287
:1413
–1419
[Abstract/Free Full Text] - Hoffenberg E, Emery L, Barriga K, et al. Clinical features of children with screening-identified evidence of celiac disease.
Pediatrics. 2004;113
:1254
–1259
[Abstract/Free Full Text] - Carnicer J, Farre C, Varea V, Vilar P, Moreno J, Artigas J. Prevalence of coeliac disease in Down's syndrome. Eur J Gastroenterol Hepatol. 2001;13 :263 –267[CrossRef][Web of Science][Medline]
- West J, Logan RF, Hill PG, et al. Seroprevalence, correlates, and characteristics of undetected coeliac disease in England.
Gut. 2003;52
:960
–965
[Abstract/Free Full Text] - Catassi C, Fabiani E, Ratsch IM, et al. The coeliac iceberg in Italy. A multicentre antigliadin antibodies screening for coeliac disease in school-age subjects. Acta Paediatr Suppl. 1996;412 :29 –35[Medline]
- Holmes GK, Prior P, Lane M, Pope D, Allan RN. Malignancy in coeliac disease: effect of a gluten free diet.
Gut. 1989;30
:333
–338
[Abstract/Free Full Text] - Fabiani E. Dietary compliance in screening-detected coeliac disease adolescents. Acta Paediatr. 1996;41 (suppl):65 –67
- Greco L, Mayer M, Ciccarelli G, Troncone RSA. Compliance to a gluten-free diet in adolescents, or "what do 300 coeliac adolescents eat every day? " Ital J Gastroenterol Hepatol. 1997;29 :305 –310[Web of Science][Medline]
- Askling J, Linet M, Gridley G, Halstensen T, Ekstrom K, Ekbom A. Cancer incidence in a population-based cohort of individuals hospitalized with celiac disease or dermatitis herpetiformis. Gastroenterology. 2002;123 :1428 –1435[CrossRef][Web of Science][Medline]
- Hermon C, Alberman E, Beral V, Swerdlow AJ. Mortality and cancer incidence in persons with Down's syndrome, their parents and siblings. Ann Hum Genet. 2001;65 :167 –176[CrossRef][Web of Science][Medline]
- Hill DA, Gridley G, Cnattingius S, et al. Mortality and cancer incidence among individuals with Down syndrome.
Arch Intern Med. 2003;163
:705
–711
[Abstract/Free Full Text] - Yang Q, Rasmussen SA, Friedman JM. Mortality associated with Down's syndrome in the USA from 1983 to 1997: a population-based study. Lancet. 2002;359 :1019 –1025[CrossRef][Web of Science][Medline]
- Gold MR, Siegel JF, Russell LB, Weinstein MC, eds. Cost Effectiveness in Health and Medicine. New York, NY: Oxford University Press; 1996
- Weinstein M, Fineberg H. Clinical Decision Analysis. Philadelphia, PA: W.B. Saunders Co; 1980
- DATA4. Decision Analysis [computer program]. Willimstown, MA: TreeAge Software, Inc; 1997
- Atkinson K, Tamakajian S, Watson W, Grejor J. Evaluation of the endomysial antibody for celiac disease: operating properties and associated cost implications in clinical practice. Can J Gastroenterol. 1997;11 :673 –677[Web of Science][Medline]
- Coeliac Disease Study Group Service Laboratory. 6/10/2003. Available at: www.uta.fi/llmama/cdserlab/. Accessed November 29, 2004
- Medicare Reimbursement. Available at: www.cms.hhs.gov/providers/pufdownload/. Accessed April 20, 2005
- Indiana Health Coverage Programs Fee Schedule. Available at: www.indianamedicaid.com/ihcp/Publications/MaxFee/fee_home.asp. Accessed June 10, 2005
- Gluten-free market. Available at: http://glutenfreemarket.com/store/enter.html. Accessed April 17, 2005
- Net grocer. Available at: www.netgrocer.com/. Accessed April 17, 2005
- Costs of care first year after diagnosis. Available at: http://progressreport.cancer.gov/doc.asp?pid=1&did=21&chid=13&coid=33&mid=vpco. Accessed April 28, 2005
- Agency for Healthcare Research and Quality. Evidence Report/Technology Assessment: Celiac Disease. Washington, DC: Agency for Healthcare Research and Quality, US Department of Health and Human Services; 2004
- Pueschel SM, Romano C, Failla P, et al. A prevalence study of celiac disease in persons with Down syndrome residing in the United States of America. Acta Paediatr. 1999;88 :953 –956[CrossRef][Web of Science][Medline]
- George EK, Mearin ML, Bouquet J, et al. High frequency of celiac disease in Down syndrome. J Pediatr. 1996;128 :555 –557[CrossRef][Web of Science][Medline]
- Friedman JM. Racial disparities in median age at death of persons with down syndrome: United States, 1968–1997. MMWR Morb Mortal Wkly Rep. 2001;50 :463 –465[Medline]
- Janicki MP, Dalton AJ, Henderson CM, Davidson PW. Mortality and morbidity among older adults with intellectual disability: health services considerations. Disabil Rehabil. 1999;21 :284 –294[CrossRef][Web of Science][Medline]
- Howdle PD, Jalal PK, Holmes GK, Houlston RS. Primary small-bowel malignancy in the UK and its association with coeliac disease.
QJM. 2003;96
:345
–353
[Abstract/Free Full Text] - Lymphoma. Available at: http://seer.cancer.gov/publications/ethnicity/lymphoma.pdf. Accessed May 2, 2005
- Hallert C, Granno C, Grant C, et al. Quality of life of adult coeliac patients treated for 10 years. Scand J Gastroenterol. 1998;33 :933 –938[CrossRef][Web of Science][Medline]
- Groot MT, Lugtenburg PJ, Hornberger J, Huijgens PC, Uyl-de Groot CA. Cost-effectiveness of rituximab (MabThera) in diffuse large B-cell lymphoma in the Netherlands. Eur J Haematol. 2005;74 :194 –202[CrossRef][Web of Science][Medline]
- Primer on cost-effectiveness analysis. Eff Clin Pract. 2003;3 :253 –255
- National Institutes of Health. Department of Health and Human Services. NIH Consensus Development Conference on Celiac Disease. Available at: http://consensus.nih.gov/2004/2004CeliacDisease118html.htm. Accessed January 15, 2005
- Harewood GC, Murray JA. Diagnostic approach to a patient with suspected celiac disease: a cost analysis. Dig Dis Sci. 2001;46 :2510 –2514[CrossRef][Web of Science][Medline]
- Mein S, Ladabaum U. Serological testing for coeliac disease in patients with symptoms of irritable bowel syndrome: a cost-effectiveness analysis. Aliment Pharmacol Ther. 2004;19 :1199 –1210[CrossRef][Web of Science][Medline]
PEDIATRICS (ISSN 1098-4275). ©2006 by the American Academy of Pediatrics
CiteULike 
Connotea 
Del.icio.us 
Digg 
Facebook 
Reddit 
Technorati 
Twitter What's this?
This article has been cited by other articles:
|
|
 |
|
  J. M. Robbins, K. S. Khan, L. M. Lisi, S. W. Robbins, S. H. Michel, and B. R. Torcato Overweight Among Young Children in the Philadelphia Health Care Centers: Incidence and Prevalence Arch Pediatr Adolesc Med, January 1, 2007; 161(1): 17 - 20. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Kemp Screening Down syndrome kids for celiac disease not cost-effective AAP News, November 1, 2006; 27(11): 2 - 2. [Full Text]
|
|
|
|
 |
|
 D. Kawatu and N. S. LeLeiko Screening for Celiac Disease in Asymptomatic Children With Down Syndrome: Cost-effectiveness of Preventing Lymphoma Pediatrics, August 1, 2006; 118(2): 816 - 817. [Full Text] [PDF]
|
|
eLetters:
Read all eLetters
- Any digestive tract tumor in Down syndrome ?
- Daniel Satge, et al.
- Pediatrics Online, 13 Nov 2006 [Full text]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
