Benefits of a Gluten-Free Diet for Asymptomatic Patients With Serologic Markers of Celiac Disease

Page created by Dolores Campos
 
CONTINUE READING
Gastroenterology 2014;147:610–617

              Benefits of a Gluten-Free Diet for Asymptomatic Patients
              With Serologic Markers of Celiac Disease
              Kalle Kurppa,1 Aku Paavola,2 Pekka Collin,2 Harri Sievänen,3 Kaija Laurila,1 Heini Huhtala,4
              Päivi Saavalainen,5 Markku Mäki,1 and Katri Kaukinen2,6
              1
               Tampere Center for Child Health Research, 4Tampere School of Health Sciences, University of Tampere and Tampere
CLINICAL AT

              University Hospital, Tampere, Finland; 2Department of Gastroenterology and Alimentary Tract Surgery, Tampere University
              Hospital and School of Medicine, University of Tampere, Tampere, Finland; 3UKK Institute, Tampere, Finland; 5Research
              Program Unit, Immunobiology and Haartman Institute, Department of Medical Genetics, University of Helsinki, Helsinki,
              Finland; 6Seinäjoki Central Hospital, Seinäjoki, Finland

                                                                                     and endomysial antibodies (EmA) are used widely in first-line
                  See Covering the Cover synopsis on page 548;                       case-finding for further endoscopic studies, but demonstration
                  see editorial on page 557.                                         of small-bowel mucosal damage still is required for the diag-
                                                                                     nosis in adults.1,6 Since celiac disease is common but difficult to
                                                                                     detect because of the heterogeneous clinical picture, a wide-
              BACKGROUND & AIMS: We investigated whether screen-
              detected and apparently asymptomatic adults with endomy-               scale screening of the whole population with noninvasive
              sial antibodies (EmA) benefit from a gluten-free diet (GFD).            serologic tests frequently has been suggested.7,8 However, the
              METHODS: We performed a prospective trial of 3031 in-                  only current treatment of the condition, a lifelong strict gluten-
              dividuals at risk for celiac disease based on screens for EmA. Of      free diet (GFD), is restrictive and difficult to maintain and thus
              148 seropositive individuals, 40 fulfilled inclusion criteria and       the positive effects of the screening are not straightforward.9
              were assigned randomly to groups placed on a GFD or gluten-            There have been some previous studies exploring the bene-
              containing diets. We evaluated ratios of small-bowel mucosal           fits of a GFD in screening-detected celiac disease patients, but
              villous height:crypt depth, serology and laboratory test results,      the results have been somewhat inconsistent, in particular with
              gastrointestinal symptom scores, physiologic well-being,               asymptomatic subjects10–14; hitherto randomized studies have
              perception of health by a visual analog scale, bone mineral            been lacking. A further problem is that wide-scale screening
              density, and body composition at baseline and after 1 year.            frequently detects seropositive subjects who are apparently
              Thereafter, the group on the gluten-containing diet started a          asymptomatic and evince only mild enteropathy or even
              GFD and was evaluated a third time; subjects in the GFD group          normal small-bowel mucosa; it has been unclear whether all
              remained on this diet. RESULTS: After 1 year on the GFD, the           such individuals in fact suffer from a true gluten-induced
              mean mucosal villous height:crypt depth values increased               clinical disease.15
              (P < .001), levels of celiac-associated antibodies decreased               The aim of this randomized trial was to establish whether
              (P < .003), and gastrointestinal symptoms improved to a greater        asymptomatic adults with positive EmA benefit from a sero-
              extent than in patients on gluten-containing diets (P ¼ .003). The     logic screening and subsequent GFD. Because celiac disease is
              GFD group also had reduced indigestion (P ¼ .006), reflux (P ¼
                                                                                     known to run in families, recruitment was executed by
              .05), and anxiety (P ¼ .025), and better health, based on the
                                                                                     screening at-risk relatives of celiac patients. By reason of its
              visual analog scale (P ¼ .017), than the gluten-containing diet
                                                                                     high specificity, EmA was selected as the inclusion criterion.1
              group. Only social function scores improved more in the gluten-
              containing diet group than in the GFD group (P ¼ .031). There
              were no differences between groups in laboratory test results,         Patients and Methods
              bone mineral density, or body composition. Most measured
              parameters improved when patients in the gluten-containing             Patients
              diet group were placed on GFDs. No subjects considered their              This study was conducted at the Department of Gastro-
              experience to be negative and most expected to remain on GFDs.         enterology and Alimentary Tract Surgery and the Tampere
              CONCLUSIONS: GFDs benefit asymptomatic EmA-positive pa-                 Center for Child Health Research, the University of Tampere,
              tients. The results support active screening of patients at risk for   and Tampere University Hospital. Relatives of celiac patients
              celiac disease. Clinicaltrials.gov no: NCT01116505.                    were recruited using newspaper advertisements and by

              Keywords: VAS; Gastrointestinal Endoscopy; Serology Test; GSRS.        Abbreviations used in this paper: BMD, bone mineral density; BMI, body
                                                                                     mass index; DGPab, antibodies to deamidated gliadin; EmA, endomysial
                                                                                     antibody; GFD, gluten-free diet; GSRS, Gastrointestinal Symptom Rating
                                                                                     Scale; IEL, intraepithelial lymphocyte; PGWB, Physiological General Well-

                  C   eliac disease is a lifelong disorder caused by inges-
                      ted cereal-derived gluten in predisposed individuals.1
              Screening studies have shown the prevalence in the European
                                                                                     Being; SF-36, Short-Form 36; TG2ab, transglutaminase 2 antibodies; VAS,
                                                                                     visual analogue scale; Vh/CrD, villous height/crypt depth ratio.

                                                                                                            © 2014 by the AGA Institute
              and North American population to be 1%–2%, and to be                                                0016-5085/$36.00
              increasing over time.2–5 Serum transglutaminase 2 (TG2ab)                            http://dx.doi.org/10.1053/j.gastro.2014.05.003
September 2014                                                                     Diagnosis of Celiac Disease Without Biopsy    611

contacting patient support groups. Altogether, 3031 in-            subjects would have remained unrecognized and continued on
dividuals volunteered for evaluation of serum celiac antibodies    a gluten-containing diet.
(Supplementary Figure 1). Exclusion criteria were a previous           All authors had access to the study data, and reviewed and
diagnosis of celiac disease, age younger than 18 years, evident    approved the final manuscript. This study has been registered
clinical symptoms, dietary gluten restriction, severe contem-      at clinicaltrials.gov; identification number NCT01116505.
porary illness or immunosuppressive medication, ongoing or
planned pregnancy, and study refusal. EmA-positive adults
(age, 18–75 y) who considered themselves asymptomatic and          Measurements
had no exclusion criteria were invited to participate in a             Serology and genetics. IgA class EmA were determined
further prospective study. The symptoms were interviewed           by an indirect immunofluorescence method using human
systemically with similar questions each time by a study           umbilical cord as substrate.16 A serum dilution of 1:5 was
physician and asymptomatic patients were defined as those           considered positive15,16 and further diluted to 1:50–1:4000.

                                                                                                                                        CLINICAL AT
with an absence of abdominal pain (>3 pain episodes over at        IgA-class TG2ab (Celikey; Phadia, Freiburg, Germany) and
least 3 months interfering with function), constipation (
612   Kurppa et al                                                                                       Gastroenterology Vol. 147, No. 3

              X-ray absorptiometry (Lunar Prodigy Advance, GE Healthcare,           Table 1.Baseline Characteristics of the Study Patients
              Waukesha, WI). BMD values were expressed as T-scores and as
              age- and sex-matched Z-scores. Changes in BMD were calcu-                                                 Gluten group        GFD group
              lated as g/cm2. BMI was determined as weight (kg)/height                        Characteristic              (n ¼ 20)           (n ¼ 20)
              (m)2; a value less than 18.5 was considered underweight,
                                                                                    Age, y
              18.5–24.9 was considered normal, 25.0–29.9 was considered
                                                                                       Median                                42                42
              overweight, and 30.0 or higher was considered obese.                     Range                                23–62             21–74
                 Small-bowel mucosal morphology and inflamma-                        Female sex, n (%)                       5 (25)            9 (45)
              tion. At each visit (at baseline, after 1 year, and after 2 years),   Hypothyroidism, n (%)                   2 (10)            1 (5)
              an upper gastrointestinal endoscopy with 6 biopsy specimens           Other chronic condition, n (%)a         7 (35)            7 (35)
              taken from the duodenum was performed and the samples                 Osteoporotic fracture, n (%)            0 (0)             0 (0)
              were coded and stored. Upon completion of the study (after the        Infertility or frequent                 1 (20)            1 (11)
CLINICAL AT

              2-year visit) the biopsy results were read and villous height             miscarriages, n (%)b
              crypt/depth ratio (Vh/CrD) was measured, and the density of           Age at menarche, yb
              CD3þ and gdþ intraepithelial lymphocytes (IEL) were counted              Median                                13                13
                                                                                       Range                                13–15             9–14
              as previously described.22 Reference values were as follows:
              Vh/CrD greater than 2.0, CD3þ IEL less than 37 cells/mm, and
              gdþ IEL less than 4.3 cells/mm.22                                     a
                                                                                        Asthma, endometriosis, diverticulosis, allergies.
                  Other evaluations. At each study visit, the subjects’             b
                                                                                        Females.
              health and medications were registered. After 1 year on a GFD
              they were asked about their dietary adherence, difficulties in
              maintaining the diet, and willingness to continue the diet in the
              future. Thereafter, they also were asked what they now thought        HLA Typing and Serology
              about their celiac disease screening.                                     All subjects had the celiac disease–associated HLA DQ2
                                                                                    or DQ8. The median EmA titer was 1:200 (range,
              Statistical Analyses                                                  1:5–1:2000) in the GFD and 1:100 (1:5–1:4000) in the
                  The primary outcome of the study defined a priori was a
                                                                                    gluten group. On intervention, the titers decreased signifi-
              change in the Vh/CrD. Previous results have shown a standard          cantly in favor of the GFD group (Figure 1A). The titers also
              deviation of 0.5 for Vh/CrD and a similar change in this value        subsequently decreased in all but 2 patients in the gluten
              has been considered clinically relevant.15,23 Consequently, to        group when they started a GFD. At baseline, TG2ab was
              provide 80% statistical power at a significance level of 0.05, 17      positive in 88% of the EmA-positive participants. All 5
              patients would be needed. Based on an estimated attrition rate        subjects who were negative for TG2ab had EmA titers of
              of 15%, we enrolled 20 subjects in each group. The chi-square         1:5–1:50. Similar to EmA, there was a significant decrease in
              test in cross-tabulations and a 2-sided Student t test or the         TG2ab in favor of GFD (Figure 1B). DGPab was positive in
              Mann–Whitney U test were used to compare differences be-              80% of the subjects, with mean values of 69  62 U in the
              tween the groups, and a paired t test or the Wilcoxon signed-         GFD group and 70  58 U in the gluten group, and changes
              rank test as used to compare changes within a group. A                after 1 year of -54  59 U and 11  38 U, respectively
              P value less than .05 was considered significant. If a subject         (P < .001).
              failed to answer 1 item in the questionnaire the missing answer
              was replaced by the median value of the other scores. If 2 or
              more items were missing the questionnaire was rejected. An-           Gastrointestinal Symptoms and Quality of Life
              alyses were conducted on an intention-to-treat principle. Cor-            The baseline mean GSRS total score was 1.8  0.6 in the
              rections for multiple comparisons were not made here because          GFD group and 1.7  0.6 in the gluten group. After inter-
              we had one predefined primary outcome and the rest were                vention, the total, indigestion, and reflux symptoms were
              secondary outcomes; in addition, no subgroup analyses were            reduced significantly in the GFD group (Figure 2A). Subse-
              performed.24 All analyses were conducted in cooperation with          quently, the total score (P ¼ .049) and indigestion
              an experienced statistician.                                          (P ¼ .016) also decreased within the gluten group after
                                                                                    starting a GFD. The PGWB total score was 112.2  12.0 in
                                                                                    the GFD group and 111.3  11.0 in the gluten group at
              Results                                                               baseline, and anxiety was alleviated in the GFD group
                  The randomized study groups were comparable with                  (Figure 2B). On intervention, the SF-36 social functioning
              respect to age, sex, medical history, and associated condi-           was reduced in the gluten group (Figure 2C). Perception of
              tions (Table 1). One subject in the gluten group started a            current health as evaluated by the VAS improved in the GFD
              GFD after randomization and 2 patients refused the 2-year             group (Figure 1D).
              endoscopy. None of the patients in the GFD group was
              willing to restart gluten, and none was withdrawn as a
              result of major symptoms or complications. Because sub-               Laboratory Parameters
              jects in the GFD group already had adhered to the diet for               The mean laboratory values were within the reference
              1 year at the time of the second endoscopy, it was consid-            range and no significant differences were observed in the
              ered unnecessary or even unethical to perform a biopsy for            changes between the groups during the intervention
              a third time after 2 years on the diet.                               (Table 2). Although there was no difference between the
September 2014                                                                    Diagnosis of Celiac Disease Without Biopsy   613

                                                                                                                                     CLINICAL AT
Figure 1. (A and B) Serum EmA and TG2ab antibodies, (C)
VAS, and (D) small-bowel mucosal Vh/CrD. The results are
shown at baseline and after 1 and 2 years of study enroll-
ment. The solid line denotes the GFD group on a gluten-free
diet and the dashed line shows the gluten group on a gluten-
containing (from baseline to 1 year) or on a gluten-free diet
(from 1 to 2 years). The dotted line denotes (A and B) upper or
(D) lower reference limits for normal. The values are              Figure 2. Absolute differences between the study groups in
expressed as the (A) unadjusted median and quartiles or            changes in the (A) GSRS, (B) PGWB, and (C) SF-36 scores.
(B–D) means and 95% confidence intervals. There was a               The differences in changes between the intervention group
significant difference in the changes between the randomi-          (GFD) and the normal diet group (gluten) are shown after 1
zation groups after 1 year in EmA (P ¼ .003) and TG2ab             year of study enrollment and are expressed as means and
(P ¼ .003), in the perception of health by VAS (P ¼ .017), and     95% confidence intervals. On the GSRS, higher scores
in the Vh/CrD (P < .001). On the VAS, the scale ranges from        denote more severe gastrointestinal symptoms, on the
0 to 100 mm and higher values indicate better subjective           PGWB higher scores denote better self-perceived quality of
perception of current health. In Vh/CrD ratios, greater than 2.0   life, and on the SF-36 higher scores denote better health
indicates normal small-bowel mucosal morphology.                   status and quality of life. Values in each subdimension scores
                                                                   are calculated as a mean of the relevant items. EP, emotional
                                                                   problems; PP, physical problems.
groups, within the GFD group the red blood cell folate
(P < .001) and vitamin B12 values improved (P < .001); this
                                                                   in the gluten group; no subject was underweight but 55%
also was seen later in the gluten group upon adaption of a
                                                                   were overweight or obese. There were no differences in BMI
GFD (P ¼ .005 and P ¼ .018, respectively).
                                                                   changes during the intervention (Table 3). The mean body
                                                                   fat percentage was 34.0  8.9 in the GFD group and 28.9 
Bone Mineral Density and Body Composition                          8.2 in the gluten group, and changes in body fat percentage
    In the GFD group the baseline mean T-score for the             did not significantly diverge between the groups (Table 3).
lumbar spine was -0.2  1.5, and for the femur neck was
-0.5  1.2; the corresponding Z-scores were -0.1  1.5 and
-0.2  1.0, respectively. In the gluten group the T-scores         Small-Bowel Mucosal Morphology and
were -0.2  1.7 and -0.3  1.0, and the Z-scores were -0.3        Inflammation
1.7 and -0.1  1.0, respectively. Differences in BMD changes         At the end of the study (after 2 years) the small-bowel
between the groups were not significant (Table 3). The              mucosal biopsy results were opened. At baseline, the
mean BMI was 27.0  6.8 in the GFD group and 26.4  3.7            mean Vh/CrD was 1.0  0.9 in the GFD group and 0.8  0.8
614   Kurppa et al                                                                                    Gastroenterology Vol. 147, No. 3

              Table 2.Laboratory Parameters at Baseline and After 1 Year         Table 3.Body Mass Index, Total Body Fat, BMD, and Density
                      on Study                                                           of Small-Bowel Mucosal IELs at Baseline and After
                                                                                         1 Year on Study
                                Gluten group         GFD group        P valuea
                                                                                                   Gluten group         GFD group         P valuea
              Blood hemoglobin level, g/dL
                Baseline           14.3  1.4         14.4    1.6               Body mass index, kg/m2
                Change             -0.2  0.6         -0.2    0.7                 Baseline          26.4  3.7           27.0  6.8
                Difference       0.0 (-0.4 to 0.4)                      .902       Change            -0.3  1.0            0.0  1.2
              Plasma albumin level, g/dL                                           Difference      0.3 (-0.5 to 1.0)                         .451
                Baseline             4.1  0.3         4.0    0.4               Total body fat, %
                Change               0.2  0.3         0.2    0.4                 Baseline          28.9  8.2           34.0  8.9
                Difference       0.0 (-0.2 to 0.2)                      .859       Change            -0.6  2.4           -1.2  3.4
CLINICAL AT

              Serum ionized calcium level, mmol/L                                  Difference      -0.5 (-2.4 to 1.4)                        .600
                Baseline           1.28  0.03        1.26    0.03              Lumbar spine BMD, g/cm2
                Change            -0.01  0.03        0.00    0.04                Baseline          1.17  0.21          1.17  0.19
                Difference       0.00 (-0.02 to 0.03)                   .687       Change           -0.01  0.03          0.00  0.02
              Plasma intact parathormone level, pmol/L                             Difference      0.01 (-0.01 to 0.02)                      .338
                Baseline             5.4  2.0         4.7    1.8               Femur neck BMD, g/cm2
                Change             -0.3  1.1         -0.3    0.9                 Baseline          1.00  0.12          0.97  0.14
                Difference       0.0 (-0.7 to 0.6)                      .916       Change           -0.01  0.03          0.00  0.02
              Serum total iron level, mmol/L                                       Difference      0.01 (-0.01 to 0.03)                      .182
                Baseline           17.3  5.7         20.0    8.6               Density of mucosal CD3þ IELs, cells/mm epithelium
                Change               2.8  6.8         0.3    7.2                 Baseline          78.5  38.1          75.9  30.2
                Difference       -2.5 (-7.0 to 2.1)                     .288       Change            -0.4  45.2         -12.9  34.1
              Red blood cell folate level, nmol/L                                  Difference      -12.5 (-39.5 to 14.4)                     .351
                Baseline            497  193         477     187               Density of mucosal gdþ IELs, cells/mm epithelium
                Change              183  215         300     260                 Baseline          29.8  19.3          24.7  12.9
                Difference       117 (-39 to 272)                       .136       Change            -2.0  20.6          -1.7  10.1
              Serum vitamin B12 level, pmol/L                                      Difference      0.4 (-10.4 to 11.2)                       .945
                Baseline            366  108         316     72
                Change                18  63           45    54
                Difference       27 (-11 to 65)                         .158     NOTE. The values are expressed as means  SD, except
              Serum alanine amino transferase level, U/L                         differences in the changes between the groups, which are
                Baseline           28.8  11.6        32.2    20.0              expressed as means with the 95% confidence interval.
                                                                                 a
                Change               0.2  10.4       -5.1    13.9                Calculated with the Student t test.
                Difference       -5.2 (-13.2 to 2.8)                    .196

                                                                                 by definition asymptomatic, subjects at baseline and after
              NOTE. The values are expressed as means  SD, except               being on a GFD.
              differences in the changes between the groups, which are
              expressed as means with the 95% confidence interval.
              a
                Calculated with the Student t test.                              Other Clinical Evaluations
                                                                                     After 2 years, when the trial was completed, 92% of the
                                                                                 subjects reported adherence to the GFD, and 8% reported
              in the gluten group. On intervention, a significant                 dietary transgressions; 85% expected to remain on the diet
              improvement in Vh/CrD was seen after 1 year on a GFD               in the future; 15% were not sure. Maintaining the GFD was
              (Figure 1D), and the ratio also increased within the gluten        considered “easy” by 5%, “goes by itself” by 67%, and
              group when they started the diet. At baseline, mucosal             “difficult” by 13%; 15% could not say. Finally, 58% expe-
              CD3þ IELs were greater than the reference value in 90%             rienced their celiac disease screening and diagnosis as
              and dgþ IELs in all but one subject; differences in IEL            positive or very positive, 42% were indifferent, and none
              changes between the groups were not significant (Table 3).          were negative.
              The Vh/CrD was considered normal (>2.0) in 2 cases in
              both groups at baseline (Table 4). Despite having morpho-
              logically normal small-bowel mucosal villi, these 4 subjects       Discussion
              showed mucosal inflammation as measured by increased                    The results of this randomized study showed that
              IELs; none of the study participants had completely normal         screen-detected and even apparently asymptomatic EmA-
              mucosa at baseline. The 2 subjects with normal Vh/CrD in           positive patients benefit from a GFD as measured by
              the GFD group evinced beneficial histologic, serologic, and         extensive clinical, serologic, and histologic parameters.
              clinical responses comparable with those with evident              There was some previous evidence indicating improvement
              villous damage. Results in the gluten group were inconsis-         in the well being of screen-detected celiac disease patients
              tent because the other subject experienced unfavorable re-         on a GFD,10–14 and we showed this now by a vigorous
              sponses while the other beneficial responses while on               randomized approach. Here, the greatest differences be-
              gluten (Table 4). Table 4 also shows examples of the               tween the groups on dietary intervention could be seen in
              symptom-specific GSRS total, diarrhea, and pain scores in 2,        serology and in mucosal morphology. For ethical reasons no
September 2014                                                                    Diagnosis of Celiac Disease Without Biopsy   615

Table 4.Data on the 4 Subjects With Positive EmA but              GSRS scores on GFD, suggest that the patients may in fact
        Normal (>2.0) Vh/CrD at Baseline                          have accepted mild symptoms as normal and recognized
                                                                  them as abnormal only later when on the diet. It is likely
                     GFD group               Gluten group
                                                                  that the changes would have been more evident in the
                 Patient 1   Patient 2   Patient 1   Patient 2    symptomatic subjects who were excluded here but
                                                                  commonly are found by active screening.11,14 In clinical
Age, y              30          36          48          34        settings many screen-detected patients also may suffer from
Sex               Female      Female      Female        Male      atypical symptoms not being recognized as celiac disease.13
EmA, titer                                                            Until now, active screening of celiac disease has been a
  Baseline         1:500        1:5         1:50        1:5
  After 1 year     1:100      Negative     1:100      Negative
                                                                  subject of controversy. Because of the high prevalence and
TG2ab, U/L                                                        ambiguous clinical picture and existence of effective treat-

                                                                                                                                     CLINICAL AT
  Baseline          84.1         4.7         8.2          5.6     ment, screening of at-risk groups or even the whole popu-
  After 1 year        7.6        0.0        19.5          0.0     lation has been suggested.7,8,13,25 Nevertheless, treatment
DGPab, U/L                                                        benefit in apparently asymptomatic patients has remained
  Baseline         125          45          16          10        unsolved. Notwithstanding the known positive effects of a
  After 1 year      24           8          65          12
                                                                  GFD, it is expensive and socially isolating and may be
Vh/CrD
  Baseline            3.4        2.6         3.1          2.2
                                                                  detrimental to quality of life. Furthermore, availability of the
  After 1 year        2.8        3.5         0.3          2.7     products is limited and they may have lower nutritional
CD3þIEL,a cells/mm epithelium                                     value and palatability than their gluten-containing coun-
  Baseline          79          95         173          81        terparts. Here, indeed, the SF-36 social functioning score
  After 1 year      76          56          92          72        decreased in the GFD group. Then again, besides objective
GSRS,b total score                                                serologic and histologic parameters, self-perceived GSRS
  Baseline            2.1        1.9         2.5          1.8
                                                                  scores and anxiety also improved and the majority of sub-
  After 1 year        1.5        1.0         2.9          2.0
GSRS, diarrhea score                                              jects showed excellent dietary adherence and were willing
  Baseline            1.3        2.0         1.7          1.7     to continue on the diet, indicating that it was not considered
  After 1 year        1.3        1.0         3.0          2.0     harmful. There also has been concern that obese celiac pa-
GSRS, pain score                                                  tients might gain more weight on a GFD26; however, the
  Baseline            2.7        1.7         2.0          1.3     present and other recent results do not support this
  After 1 year        1.7        1.0         1.3          1.7     concept.27,28 Furthermore, although not as apparent here,
PGWB,c total score
  Baseline         117         102         104         102
                                                                  there is evidence that screen-detected patients may have a
  After 1 year     120         122          83         107        decreased quality of life and BMD that improves on a
VAS,d mm                                                          GFD.10–13 Nevertheless, besides the decrease in social
  Baseline          91          81          72          91        functioning seen here, in some screen-detected patients the
  After 1 year      94          96          46          89        celiac disease diagnosis may lead to increased anxiety and
                                                                  health concerns.10 Given that unrecognized celiac disease
NOTE. Results are shown at baseline and after 1 year on a
                                                                  may not necessarily increase the risk of malignancy or
GFD (GFD group) or on a normal, gluten-containing diet            mortality,29–31 the consequences of possible screening must
(gluten group).                                                   be weighed carefully in advance on an individual basis.
a
  Density of intraepithelial lymphocytes; reference value < 37    Furthermore, because the present study was conducted in
cells/mm epithelium.                                              individuals belonging to an at-risk group, the question of
b
  Higher scores indicate more severe self-perceived gastro-       population-based mass screening remains a subject for
intestinal symptoms.                                              future studies.
c
  Higher scores indicate better self-perceived health-related
quality of life.                                                      As a result of increasing celiac disease screening, a
d
  Higher scores indicate better self-perceived general health.    substantial number of seropositive individuals manifest
                                                                  with only mild enteropathy (ie, Marsh grades I–II).15 Thus
                                                                  far, strict criteria with the demonstration of villous atrophy
                                                                  (Marsh III) have been necessary to ensure a reliable diag-
subjects with evident symptoms were randomized and, in            nosis of this lifelong and restrictive disorder. However, the
fact, the mean baseline quality of life was even better than      modern biomarkers, particularly EmA, have shown almost
previously observed in healthy controls.10 Likewise, the          100% specificity for celiac disease.1 Furthermore, we
mean laboratory parameters and BMD were within normal             previously showed that most false-positive cases with
range; one cannot expect major improvements in in-                initially morphologically normal villi (Marsh I–II) eventu-
dividuals at such an early stage in celiac disease. Nonethe-      ally will develop mucosal atrophy if they continue on
less, in favor of a GFD, there were also significant differences   gluten.32 Moreover, a randomized trial showed that these
in the GSRS and quality-of-life scores. This randomized           EmA-positive subjects already benefit from a GFD before
approach showed that subjects who thought they were               villous atrophy (the end stage of the histologic damage)
asymptomatic experienced improvement in several objec-            develops.15 Accordingly, the latest diagnostic criteria of
tive disease scores upon adopting a GFD. These findings, as        celiac disease have begun to accept milder forms of en-
exemplified in Table 4 by a marked decrease in most of the         teropathy as diagnostic in seropositive patients.17 These
616   Kurppa et al                                                                               Gastroenterology Vol. 147, No. 3

              findings suggest that the hitherto used grouped classifica-        Supplementary Material
              tions are no longer optimal in the diagnosis of celiac dis-      Note: To access the supplementary material accompanying
              ease. Nevertheless, because serology and symptoms have           this article, visit the online version of Gastroenterology at
              been shown to be unreliable markers for the mucosal              www.gastrojournal.org, and at http://dx.doi.org/10.1053/
              improvement and dietary adherence in celiac disease,6 we         j.gastro.2014.05.003.
              considered it necessary to use histology as the primary
              outcome in this study; this is likely even more important in
              screen-detected subjects with mild clinical presentation. To
              detect even minor changes in morphology, validated and               References
              quantitative Vh/CrD was used in this study instead of             1. Di Sabatino A, Corazza GR. Coeliac disease. Lancet
              grouped classification.33                                             2009;373:1480–1493.
                                                                                2. Mäki M, Mustalahti K, Kokkonen J, et al. Prevalence of
CLINICAL AT

                  In this study, all EmA-positive subjects had celiac dis-
              ease–associated HLA and all but 1 subject had an increased           celiac disease among children in Finland. N Engl J Med
              dgþ IEL count, which is further proof of the presence of             2003;348:2517–2524.
              celiac disease irrespective of the degree of villous dam-         3. Lohi S, Mustalahti K, Kaukinen K, et al. Increasing
              age.22,34,35 Furthermore, the seropositive cases with clear          prevalence of coeliac disease over time. Aliment Phar-
              mucosal inflammation but normal villous morphology                    macol Ther 2007;26:1217–1225.
              showed a beneficial response to GFD in most of the                 4. Rubio-Tapia A, Kyle RA, Kaplan EL, et al. Increased
              measured parameters. Nevertheless, it still remains unclear          prevalence and mortality in undiagnosed celiac disease.
                                                                                   Gastroenterology 2009;137:88–93.
              if all EmA-positive, screen-detected individuals should be
              treated. None of the participants in this study or in our         5. Catassi C, Kryszak D, Bhatti B, et al. Natural history of
                                                                                   celiac disease autoimmunity in a USA cohort followed
              previous randomized study investigating the role of EmA in
                                                                                   since 1974. Ann Med 2010;42:530–538.
              diagnostics15 had a completely normal biopsy (Marsh 0) at
                                                                                6. Rubio-Tapia A, Hill ID, Kelly CP, et al. ACG clinical
              baseline and further prospective studies are needed to
                                                                                   guidelines: diagnosis and management of celiac disease.
              decipher whether these subjects truly benefit from a GFD. If
                                                                                   Am J Gastroenterol 2013;108:656–676.
              so, even the universal need for a biopsy in EmA-positive
                                                                                7. Collin P. Should adults be screened for celiac disease?
              adults might be questioned. Such studies, however, may
                                                                                   What are the benefits and harms of screening? Gastro-
              be difficult to execute because a very large at-risk cohort
                                                                                   enterology 2005;128:S104–S108.
              likely should be screened to achieve sufficient power.
                                                                                8. Fasano A. European and North American populations
                  There were certain limitations to the present study. It
                                                                                   should be screened for coeliac disease. Gut 2003;52:
              was impossible to blind subjects to a GFD and some placebo
                                                                                   168–169.
              effect may have accounted for the results. Furthermore,
                                                                                9. Leffler DA, Edwards-George J, Dennis M, et al. Factors
              subjects in the gluten group might have been more sensitive
                                                                                   that influence adherence to a gluten-free diet in adults
              to possible celiac disease–associated symptoms, and we               with celiac disease. Dig Dis Sci 2008;53:1573–1581.
              cannot exclude gluten reduction at the individual level.
                                                                               10. Ukkola A, Mäki M, Kurppa K, et al. Diet improves
              Because EmA positivity was an inclusion criterion, there             perception of health and well-being in symptomatic, but
              were no EmA-negative TG2ab-positive, or entirely sero-               not asymptomatic, patients with celiac disease. Clin
              negative, patients, and the results from this study should not       Gastroenterol Hepatol 2011;9:118–123.
              be extended to apply to such individuals. Another limitation     11. Mustalahti K, Collin P, Sievänen H, et al. Osteopenia in
              was that no other indicators for bone metabolism apart               patients with clinically silent coeliac disease warrants
              from BMD, for example, vitamin D levels, were measured. It           screening. Lancet 1999;354:744–745.
              also must be borne in mind that, in general, there is good       12. Mazure R, Vazquez H, Gonzalez D, et al. Bone mineral
              knowledge of celiac disease and easy availability of gluten-         affection in asymptomatic adult patients with celiac dis-
              free products in Finland, but this might not be the case in          ease. Am J Gastroenterol 1994;89:2130–2134.
              many other countries, for example, in the United States.9,36     13. Vilppula A, Kaukinen K, Luostarinen L, et al. Clinical
              In contrast to celiac disease in general,1 in this study 65%         benefit of gluten-free diet in screen-detected older celiac
              of the randomized subjects were males, the reason being              disease patients. BMC Gastroenterol 2011;11:136.
              that a considerably higher proportion of females than males      14. Paavola A, Kurppa K, Ukkola A, et al. Gastrointestinal
              fulfilled the age exclusion criteria.                                 symptoms and quality of life in screen-detected celiac
                  In conclusion, our randomized study showed that                  disease. Dig Liver Dis 2012;44:814–818.
              apparently asymptomatic EmA-positive subjects benefit             15. Kurppa K, Collin P, Viljamaa M, et al. Diagnosing mild
              from serologic screening and a subsequent GFD. The results           enteropathy celiac disease: a randomized, controlled
              support active screening of celiac disease in at-risk groups.        clinical study. Gastroenterology 2009;136:816–823.
              However, because of the possible detrimental effects of a        16. Ladinser B, Rossipal E, Pittschieler K. Endomysium an-
              GFD in social functioning, the consequences of screening             tibodies in coeliac disease: an improved method. Gut
              must be prejudged on an individual basis. Furthermore,               1994;35:776–778.
              prospective studies are needed to unravel whether screen-        17. Husby S, Koletzko S, Korponay-Szabó IR, Mearin ML,
              detected seropositive subjects with completely normal                Phillips A, Shamir R, Troncone R, Giersiepen K, et al.
              small-bowel mucosal histology should be treated.                     EPGHAN guidelines for the diagnosis of coeliac disease
September 2014                                                                           Diagnosis of Celiac Disease Without Biopsy            617

      in children and adolescents: an evidence-based                  30. Godfrey JD, Brantner TL, Brinjikji W, et al. Morbidity and
      approach. J Ped Gastroenterol Nutr 2012;54:136–160.                 mortality among older individuals with undiagnosed ce-
18.   Hill PG, Holmes GK. Coeliac disease: a biopsy is not                liac disease. Gastroenterology 2010;139:763–769.
      always necessary for diagnosis. Aliment Pharmacol Ther          31. Lohi S, Mäki M, Montonen J, et al. Malignancies in cases
      2008;27:572–577.                                                    with screening-identified evidence of coeliac disease: a
19.   Svedlund J, Sjödin I, Dotevall G. GSRS–-a clinical rating           long-term population-based cohort study. Gut 2009;
      scale for gastrointestinal symptoms in patients with irri-          58:643–647.
      table bowel syndrome and peptic ulcer disease. Dig Dis          32. Collin P, Helin H, Mäki M, et al. Follow-up of patients
      Sci 1988;33:129–134.                                                positive in reticulin and gliadin antibody tests with normal
20.   Dupuy HJ. The Psychological General Well-Being                      small-bowel biopsy findings. Scand J Gastroenterol
      (PGWB) Index. In: Wenger NK, Mattson ME,                            1993;28:595–598.
      Furberg CD, et al., eds. Assessment of quality of life in       33. Taavela J, Koskinen O, Huhtala H, et al. Validation of

                                                                                                                                                       CLINICAL AT
      clinical trial of cardiovascular therapies. New York: Le            morphometric analyses of small-intestinal biopsy read-
      Jacq Publishing, 1984:184–188.                                      outs in celiac disease. PLoS One 2013;8:e76163.
21.   McHorney CA, Ware JE Jr, Lu JF, et al. The MOS 36-item          34. Karell K, Louka AS, Moodie SJ, et al. HLA types in celiac
      Short-Form Health Survey (SF-36): III. Tests of data                disease patients not carrying the DQA1*05-DQB1*02
      quality, scaling assumptions, and reliability across                (DQ2) heterodimer: results from the European Genetics
      diverse patient groups. Med Care 1994;32:40–66.                     Cluster on Celiac Disease. Hum Immunol 2003;64:
22.   Järvinen T, Kaukinen K, Laurila K, et al. Intraepithelial           469–477.
      lymphocytes in celiac disease. Am J Gastroenterol 2003;         35. Savilahti E, Arato A, Verkasalo M. Intestinal gamma/
      98:1332–1337.                                                       delta receptor-bearing T lymphocytes in celiac disease
23.   Kaukinen K, Peräaho M, Collin P, et al. Small-bowel                 and inflammatory bowel diseases in children. Constant
      mucosal transglutaminase 2-specific IgA deposits in                  increase in celiac disease. Pediatr Res 1990;28:
      coeliac disease without villous atrophy: a prospective              579–581.
      and randomized clinical study. Scand J Gastroenterol            36. Kurppa K, Lauronen O, Collin P, et al. Factors associated
      2005;40:564–572.                                                    with dietary adherence in celiac disease: a nationwide
24.   Perneger TV. What’s wrong with Bonferroni adjustments.              study. Digestion 2012;86:309–314.
      BMJ 1998;316:1236–1238.
25.   Katz KD, Rashtak S, Lahr BD, et al. Screening for celiac
      disease in a North American population: sequential
                                                                      Author names in bold designate shared co-first authors.
      serology and gastrointestinal symptoms. Am J Gastro-
      enterol 2011;106:1333–1339.                                     Received August 15, 2013. Accepted May 6, 2014.
26.   Dickey W, Kearney N. Overweight in celiac disease:              Reprint requests
      prevalence, clinical characteristics, and effect of a gluten-   Address requests for reprints to: Katri Kaukinen, MD, University of Tampere,
      free diet. Am J Gastroenterol 2006;101:2356–2359.               School of Medicine, FIN-33014, Tampere, Finland. e-mail: katri.kaukinen@uta.fi;
                                                                      fax: (358) 3-3551-8402.
27.   Cheng J, Brar PS, Lee AR, et al. Body mass index in
      celiac disease: beneficial effect of a gluten-free diet.         Conflicts of interest
                                                                      The authors disclose no conflicts.
      J Clin Gastroenterol 2010;44:267–271.
28.   Ukkola A, Mäki M, Kurppa K, et al. Changes in body              Funding
      mass index on a gluten-free diet in coeliac disease: a          This work was supported by the Academy of Finland Research Council for
                                                                      Health, the Competitive Research Funding of the Pirkanmaa Hospital District
      nationwide study. Eur J Intern Med 2012;23:384–388.             (Grants 9R034 and 9R018), the Sigrid Juselius Foundation, the Finnish
29.   Lohi S, Mäki M, Rissanen H, et al. Prognosis of unrecog-        Foundation for Gastroenterological Research, the Yrjö Jahnsson Foundation,
                                                                      the Finnish Medical Foundation, the Foundation for Pediatric Research and
      nized coeliac disease as regards mortality: a population-       the Finnish Celiac Society. The sponsors had no role in the contents of the
      based cohort study. Ann Med 2009;41:508–515.                    article or the decision to submit the manuscript for publication.
617.e1   Kurppa et al                                                                          Gastroenterology Vol. 147, No. 3

Supplementary Figure 1. Enrollment, randomization, and follow-up evaluation of the study patients. *Exclusion criteria of the
study were as follows: previous diagnosis of celiac disease, age younger than 18 years, evident clinical symptoms, dietary
gluten restriction, severe contemporary illness, pregnancy, immunosuppressive medication, and study refusal. †One subject
started a gluten-free diet soon after randomization, but was assigned to the gluten group according to an intention-to-treat
principle. ‡The subjects remained on a gluten-free diet.
You can also read