Predictive Value of Serology Testing in Celiac Disease
                              Presented by
       Dr. Vijay Kumar at Mt. Sinai Medical Centre, Nov. 9, 1996
          American Celiac Society/Dietary Support Coalition
                      Mystery Golden Key Conference
                         condensed by Ron Hoggan

   Dr. Kumar has been working on serological testing for celiac disease 
over the last 15 years, and was presenting after having been up all night 
trying to catch a flight from Buffalo after a severe snow storm. We were all
pleased by his dedication, as his presentation was both informative and
compelling.

   He began by asking some fundamental questions, which challenged several 
common beliefs about the diagnosis of celiac disease, beginning with:
"Can we effectively diagnose celiac disease by its clinical presentations?"

   The answer came in the data he cited from Campbell & Davidson in their 
survey of 1300+ celiac patients. The first item indicated that celiac
disease may present with a broad variety of symptoms, and are easy for
clinicians to miss. The second item indicated that the majority had visited
5 or more doctors prior to diagnosis. The third item indicated that it had
taken an average of 5 to 10 years, after initial presentation, for celiac
disease to be diagnosed.   

   Clearly, the answer to Dr. Kumar's question was that celiac disease can
not be diagnosed by its clinical presentations. What would be the value of
histological, or any other evidence, if it could be diagnosed on the basis
of clinical presentation alone? He mentioned the considerable risk of
misdiagnosis as well. While clinical presentation may help in identifying
the possibility of celiac disease, it can not, reasonably, be the sole
criteria for an accurate diagnosis. 

   "Why do we need an early diagnosis of celiac disease?" was his next 
question. The answer was twofold. First, young children may not grow and 
develop properly if they have unidentified celiac disease. Normal  
development, for many of these children, requires that they be on a 
gluten-free diet, thus recieving adequate nutrition. Second, untreated 
celiac disease is associated with a very high risk of lymphoma, and the 
gluten free diet plays a protective role against lymphoma. It reduces the
risk to almost the same as the general population, after 5 years on a strict
diet. He cautioned that less than 5 years may not yeild statistically
significant reductions in risk. Similar risk factors for malignancy are at
work in dermatitis herpetiformis.   

   Histopathology gained acceptance in the mid-sixties, to aid in the 
diagnosis of celiac disease, because of the problems with diagnosis on 
the basis of clinical presentation. This is where a biopsy is taken from 
the small intestine, and the morphology is examined to determine the 
status of villi in the patient. Slides were shown, first demonstrating 
normal morphology of healthy villi, then demonstrating total villous
atrophy, along with crypt hyperplasia, and increased density of
intraepithelial lymphocytes. These were the criteria proposed in the
mid-sixties, but such morphology alone is not diagnostic of celiac disease.
Many other diseases like parasitic infections, immune deficiency disorders,
and nutritional deficiency disorders, all look quite similar. 

At that time, ESPGAN (European Society of Gastroenterology and Nutrition) 
proposed a strict set of diagnostic criteria for celiac disease. First, 
characteristic flat or damaged villi were to be demonstrated. Second, 
after a gluten-free diet for an extended period of time,  when the patient
was feeling normal, a second biopsy should be taken, showing villi of 
normal appearance. Finally, a gluten-containing diet would be re-instituted,
and if after a period of about 6 months, if villous atrophy was demonstrated 
at biopsy, then the diagnosis of celiac disease was confirmed. 

   In many of the patients he has seen, there are poblems getting them to 
eat enough gluten to confirm the diagnosis for the third biopsy. Then 
there is the limitation that initially, patients need to be identified as 
needing the biopsy. Yet another limitation of this approach is that when a
biopsy is taken, it is removed from a small, localized area. It may miss a
region of typical celiac intestinal lesion and take the sample from a 
region showing very mild symptoms which are difficult to identify. These 
histological criteria are on a continuum, showing flat villi at one end, 
and normal villi at the other. Yet another limitation arises in the three
recent reports which identify celiac disease in patients with normal
histopathology. Pathologists apparently read them as normal biopsies. He
cited papers from 1982, 1993 and 1996. 

   Dr. Kumar used the iceburg metaphor to describe the current status of 
recognition of celiac disease, where the vast majority of celiac disease 
remains undiagnosed. We are seeing only typical celiac disease. This is 
where weight loss, diarrhea, short stature, very typical manifestations,  
lead to the biopsy, where the histology reveals characteristic damaged
villi. What remains under the surface is two large groups: 1) silent celiac
disease where the the patient presents with mild, atypical manifestations
like low levels of iron and low haemoglobin levels. Some patients only
present with aphthous ulcerations in their mouths but if they are tested,
many of them demonstrate characteristic histopathology; 2) Then you have
situations where patients who really have celiac disease, but if you test
them, they show normal histopathology. What can be done, then? 

   When it was first identified, in the 1960's that the gliadin proteins in
wheat were the toxic entities in celiac disease, many investigators in
Europe began looking for anti-gliadin antibodies. For serology to be
effective, the techniques have to be very refined and well standardized.
Otherwise, one lab reporting results may not be as good as another lab
reporting results. 

   Dr. Kumar reported on his own lab results in anti-gliadin antibody 
tests. He indicated that 5% to 10% of those demonstrating IgG anti-gliadin
antibodies do not have celiac disease. Rarely, on the other hand, are IgA 
anti-gliadin antibodies demonstrated in non-celiacs. These antibodies are 
detected by very simple methods. The consensus in the field is that IgG 
antibodies are more sensitive, but not specific, and IgA antibodies are 
more specific but less sensitive. 
   
   The next antibody identified in the literature was the antireticulin 
antibody. We don't really know what protein antigen this antibody is 
developed in response to. The name is a serological term, but it reacts to
substances called reticulin which is in the tissues surrounding the 
tubules in the kidney. There are five types of reticulin antibodies. Only 
one of these is associated with celiac disease, and this is not a very 
sensitive marker, but it is very specific for celiac disease. When a
patient goes on a g-f diet, their antibody levels will eventually disappear
and become negative.    

   The third type of antibody, reported in 1984, was the endomysial 
antibody. It is named in reference to the endomysium which is 
the lining of the muscle fibers. These, and the antireticulin antibodies 
are primarily of IgA class. In situations where patients are IgA 
deficient, they will make IgG antibodies. Patients with celiac disease 
demonstrate this antibody with virtually 100% frequency. On a gluten-free 
diet, these antibody levels become normal, or negative. They are a very 
sensitive, very specific marker of celiac disease. 

   It takes varying periods for these levels to become normal. They 
reappear very quickly on a gluten challenge. He then reported on 133 
celiac patients, 132 of whom showed IgA antiendomysium antibodies when 
eating gluten. The one who did not was IgA deficient, and was positive for
IgG endomysial antibody. IgA is the more sensitive marker. 
 
   All of the 133 patients eventually demonstrated negative levels of 
antiendomysial antibodies. When rechallenged with gluten 130 once again 
became positive for antiendomysial antibodies. Other studies indicate 
that if the other 3 were challenged with a high gluten-containing diet, 
they would have demonstrated antibodies too. 

   Another group of 31 showed abnormal biopsies, but endomysial 
antibodies were not demonstrated. On a gluten-free diet, the the villous 
morphology returned to normal, but could not be induced again on gluten 
challenge. Neither were antiendomysial antibodies detected. It appears 
that these were not cases of celiac disease. A single biopsy can be 
misleading without additional evidence. 

   Yet another group of patients in Italy, reported in 1996, demonstrated 
symptoms suggestive of celiac disease. They tested positive for 
antiendomysial antibodies, but did not display abnormalities in villous 
morphology. The question is, are these patients celiac or not? They
demonstrated typical symptoms, which resolved on the gluten-free diet, and
the endomysium antibodies also disappeared on the diet. Upon gluten 
challenge, these symptoms returned, and the high antibody levels 
returned also. The clinical manifestations, including diarrhea,  
steathorrea, and abdominal cramping, are convincing evidence that these 
are cases of celiac disease in the early stages of immune response which, 
with time, will develop typical histology of celiac disease. Follow-up 
studies have demonstrated just this. 

   There are certain HLA markers which are associated with celiac 
disease. Nine of ten celiac patients in another study did not have these
markers. That means that the genetic markers we currently associate 
with celiac disease are also problematic. Lymphocyte density was also within
the normal range. This indicates that these individuals are very early in
their immune response. He used demonstrations of increased ICAM presentation
in the same tissue, to indicate that an immune response was being mounted,
but it was at a very early stage. If allowed to continue, Dr. Kumar was
confident that the typical histopathology of celiac disease would eventually
be demonstrated. 

   He went on to show a study reported in _Acta Paediatrica_ where 
patients who demonstrated antiendomysial antibodies and normal mucosa, 
were re-biopsied three years later, and demonstrated villous atrophy. 
Quantity of gluten intake seems to be the primary variable, and where 
increased intake occurs villous atrophy follows. 

   Examples reported as recently as 1994 indicate that diagnosis may still
be slow among elderly patients, even where there is a threat to life.
This demonstrates that diagnosis of celiac disease continues to be a very 
slow process in many cases. A simple serological test could help immensely.

   Among family members of those with celiac disease, symptomatic or 
asymptomatic celiac disease is present in 10% to 15% of first degree 
relatives. This is cd associated with short stature, autoimmune disorders
and cardiac abnormalities. A simple serological test should be 
considered for family members.

   Dr. Kumar presented a compelling case for the value of serological 
screening for celiac disease. He indicated that the biopsy remains the 
gold standard for diagnosis, but I walked away from his presentation 
convinced that the serological testing offers earlier diagnosis, and thus 
improved health for celiac patients. We may be seeing the emergence of a 
new gold standard in the diagnosis of celiac disease; one which may 
reduce risk, be more reliable than the biopsy, and it is less invasive.