Intractable Diarrhea With Tufting Enteropathy: A Favorable Outcome Is Possible

Abstract

Background and Objective: 

Tufting enteropathy (TE) is a congenital abnormality of intestinal mucosa development characterized by severe intestinal failure requiring parenteral nutrition (PN) and, in some cases, small bowel transplantation. A few patients have had a more favorable outcome. The objective of this study was to evaluate possible correlations between histological lesion severity in duodenal biopsies and clinical outcomes in children with TE.

Patients and Methods: 

We retrospectively reviewed the records of patients diagnosed with TE between 1993 and 2003 at our institution based on intractable neonatal-onset diarrhea with prolonged dependence on PN and duodenal biopsy findings of villous atrophy, epithelial dysplasia with enterocyte dedifferentiation and disorganization (tufting) of the surface epithelium, and crypt abnormalities. The histological lesions were assessed semiquantitatively and compared with the clinical outcomes including dependence on PN.

Results: 

Seven children, all from consanguineous parents, were studied for a median of 6.5 years. Three were permanently weaned off PN and experienced normal growth without nutritional assistance. Initial biopsies in all 3 children showed severe diffuse histological lesions. At weaning off PN, 2 of these 3 patients had persistent, although less diffuse, histological lesions.

Conclusions: 

Progressive weaning off PN is possible in some children with TE. In our experience, this favorable outcome was not predicted by histological lesion severity, although the lesions improved in some patients. New biomarkers for identifying the histological lesions and predicting the outcome would be useful.

Tufting enteropathy (TE), also known as intestinal epithelial dysplasia, is 1 of the causes of intractable diarrhea in neonates ^(1–3). First described in 1994 by Reifen et al ⁽⁴⁾, TE is a congenital abnormality of intestinal mucosa development characterized by a specific histological pattern of persistent villous atrophy and epithelial dysplasia. It is probably inherited on an autosomal recessive basis, and mutations in the EpCAM gene were identified recently in some patients ^(5,6). The prevalence of the disease can be estimated at around 1/50,000 to 100,000 live births in western Europe ⁽⁵⁾. TE causes severe malabsorption that usually requires lifelong parenteral nutrition (PN). Intestinal transplantation has been used in a few PN-dependent patients ^(7,8). A favorable outcome with definitive weaning from PN has been reported ⁽⁹⁾, suggesting genetic and pathophysiological differences across patients.

During an 11-year period (1993–2003), 7 patients with TE were managed at our institution, including 3 patients who were permanently weaned off PN. The aim of this retrospective study was to evaluate associations between histological lesion severity in duodenal biopsies and clinical outcomes in our patients.

PATIENTS AND METHODS

Patients

We included all of the children admitted to our institution between 1993 and 2003 and diagnosed with TE based on ⁽⁴⁾ intractable neonatal-onset diarrhea with long-term dependence on PN combined with a characteristic histological pattern identified by duodenal biopsy. This pattern was defined as varying degrees of villous atrophy (from minimal to subtotal); epithelial dysplasia with enterocyte dedifferentiation and disorganization, and rounding of the apical plasma membrane, giving the surface epithelium a “tufted” appearance; crypt abnormalities (hyperplasia, dilatation, and branching); no evidence of inflammatory changes in the lamina propria; fewer than 15% intraepithelial lymphocytes; and Paneth cell hyperplasia.

Exclusion criteria were as follows: follow-up shorter than 3 years, congenital diseases of digestion and absorption other than TE, and immunological or other constitutional cause of intractable diarrhea (microvillous inclusion disease or syndromic enteropathy).

Data Collection

The following data were abstracted from the medical records of each patient: family history, age at onset and characteristics of diarrhea, changes in diarrhea over time, weight gain and statural growth, degree of dependence on PN defined as the percentage of theoretical age–related energy needs supplied by PN, complications of PN, ages at which duodenal biopsy was performed, and histological findings.

The duodenal biopsy specimens were obtained during endoscopy and fixed in formalin or Bouin fixative. Sections were stained with hematoxylin-eosin-saffron and periodic acid-Schiff (PAS). All of the biopsies from each patient were examined retrospectively by the same pathologist (A.C.), who assessed the histological abnormalities semiquantitatively as mild, moderate, or severe. Histological findings over time were compared and correlated with the degree of PN dependence.

RESULTS

Initial Clinical Data

Seven children (4 boys and 3 girls) were enrolled and followed up for a median of 84 months (57 to 171 months) (Table 1). All of the patients were born to parents who were consanguineous in the second degree. Gestation ranged from 37 to 40 weeks and was uneventful in all of the cases. More specifically, no history of hydramnios was noted. Birth weight ranged from 2410 to 3430 g (median 3260 g). In 3 patients, a family history of neonatal death caused by intractable diarrhea was found: patients 1 and 6 each had a sibling who died and patient 7 had a second cousin who died.

TABLE 1: Clinical features in the 7 patients

Diarrhea onset was within 1 week after birth in the first week of life for 5 patients and within 1 month after birth in 2 patients (1 and 4). Mucus was present in the stools of all 7 patients. At diarrhea onset, 4 patients were breast-fed and 3 were fed with standard infant formula. At inclusion, all of the patients presented 8 to 10 stools per day. The precise daily stool output was not known. Bouts of vomiting starting at birth were noted in 4 patients.

None of the patients had any clinically obvious phenotypic abnormalities such as choanal atresia, anal imperforation, or skeletal abnormalities. None of the patients had received immunosuppressive treatment.

Nutritional Care and Outcome

Total PN was started between 20 days and 6 months of age (median 3 months). With total PN, the vomiting stopped in the 4 patients with this feature, but the diarrhea persisted in all 7 patients; however, the diarrhea improved in 2 patients (6 and 7). After a median time on total PN of 59 days (range 42–105 days), oral or enteral feeding was reintroduced progressively in all of the patients.

With PN, weight gain and statural growth were normal for the 7 patients (Table 2). Of the 7 children, 4 remained dependent on PN throughout follow-up, although the degree of dependency varied across patients. At last follow-up (age range 4.9–7.2 years), PN provided 55% to 90% of the theoretical age–specific energy needs in these 4 patients (Table 1). Furthermore, 3 of these patients also required continuous enteral feeding at night. The remaining 3 children were successfully weaned off PN at 3, 3.5, and 12 years old, respectively (see description below). Diarrhea improved in all patients over time except in patient 4 (Table 1).

TABLE 2: Nutritional status, PN dependence, and histology correlation over time

No complications of PN were seen in 6 of the 7 children. The remaining patient experienced extensive superior vena cava thrombosis at the time of weaning. During follow-up, 1 child (3) experienced recurrent episodes of acute pancreatitis, for which no surgical or medical causes were found. In addition, patient 6 had bronchial disease of undetermined etiology. None of the patients had any ophthalmological abnormalities such as punctate keratitis detected during follow-up.

Histological Data

The median of duodenal biopsies during follow-up was 6 per patient (range 3–11). Age at first biopsy ranged from 3 to 7 months (median 5.5 months). For all of the patients, 2 or 3 bioptic controls were performed in the first year of life and then approximately every 2 years during the duration of the PN.

Villous atrophy was found consistently. Severity varied across patients from minimal to subtotal. In a given patient, the severity of villous atrophy varied over time and these variations showed no obvious correlation to the clinical course.

Surface enterocyte disorganization was a consistent finding. Epithelial dysplasia with focal crowding of closely packed enterocytes and rounding of the apical plasma membrane produced a teardrop appearance (Fig. 1A and B). These epithelial abnormalities varied over time and were often absent from the first biopsy. Thus, tufts with focal crowding were seen in biopsies taken in patients between 3 and 11 months old (median 8 months) in 6 patients; in the remaining patient (5), tufts occurred only at 5.5 years old. Surface enterocyte disorganization was heterogeneous and varied both across patients and across biopsies in each patient. Two patients (1 and 7) had extensive surface epithelium lesions before 1 year old (Fig. 1A) that persisted until 3 years old and then became more focal (Fig. 1C). In the other 5 patients, the epithelial lesions were chiefly focal and showed little change over time (Fig. 1D and E).

FIGURE 1: A–C, Patient 1. Biopsies in the first year of life: extensive disorganization of the surface enterocytes with focal crowding and pseudocystic crypts (A) (hematoxylin & eosin stain, original magnification ×10), typical tuft (B) (hematoxylin & eosin stain, original magnification ×40). Less extensive lesions of the surface epithelium with improvement in the appearance of the crypts before definitive weaning off parenteral nutrition (C) (hematoxylin & eosin stain, original magnification ×20). D–F, Patient 3. Tufts (D) (hematoxylin & eosin stain, original magnification ×40). The epithelial surface damage is less diffuse and crypt branching is seen (E) (hematoxylin & eosin stain, original magnification ×5), Paneth cell hyperplasia (F) (hematoxylin & eosin stain, original magnification ×10).

Dilated pseudocystic crypts were an early finding in all of the patients, arising before or at the same time as the epithelial lesions. These crypt abnormalities decreased or even disappeared with time in all of the patients (Fig. 1A). Crypt branching was seen concomitantly with the epithelial lesions and was associated with epithelial lesion severity (Fig. 1E), being mild to moderate in patients with focal epithelial lesions and severe in patients with diffuse epithelial lesions. The number of mitoses was normal or increased (up to 3 per crypt) and changed over time.

In all of the patients, the brush border was normal by PAS staining. The percentage of intraepithelial lymphocytes was normal (<15%) or decreased (<5%) in all of the biopsies from all of the patients. In all of the patients, regardless of whether the histological lesions were diffuse or focal, the proportion of inflammatory cells in the lamina propria was normal or low. In addition, all of the patients had Paneth cell hyperplasia on at least 1 biopsy. This change persisted but was moderate (Fig. 1F). None of the patients had Brunner gland abnormalities.

Clinical and Histological Characteristics of the Patients Weaned Off PN

Three patients were successfully weaned off PN, at 12, 3.5, and 3 years of age, respectively. All 3 patients were of Moroccan descent (Table 1). Without nutritional support, their weight gain and statural growth were normal during the median follow-up of 7 years (range 4–12 years) (Table 2). Two patients had normal stools, but the remaining patient (1) continued to experience episodes of diarrhea with vitamin malabsorption when on a normal diet. This patient received a supplementation of vitamin B₁₂.

Initially, the biopsies in 2 of the 3 weaned patients (1 and 7) showed extensive disorganization of the surface epithelium with numerous tufts, whereas the lesions were more focal in the third patient (6). Over time, the histological abnormalities improved in all 3 patients. In patients 1 and 7, the tufts persisted but had a more focal distribution at the time of definitive weaning. In patient 6, a biopsy performed a few months before definitive weaning showed minimal histological lesions with no tufts.

Crypt branching and/or crypt pseudocysts improved over time in 2 of these 3 patients. The remaining patient (1) had persistent epithelial lesions involving the crypt lumen before definitive PN weaning. Four years after PN weaning, a duodenal biopsy from this patient showed persistent, mild, focal epithelial lesions. In the remaining 2 patients, no biopsies were taken at a distance from weaning off PN (Table 3).

TABLE 3: Patients definitively weaned off PN: changes in histological findings

In comparison, the 4 patients who were not weaned off PN initially had epithelial lesions with typical tufts. These lesions initially seemed less severe than for the weaned patients. These abnormalities were focally distributed in all 4 patients and showed little change over time (Table 2). No frank improvement of the histological lesions was noted during the evolution.

DISCUSSION

TE is a rare inherited disease that usually causes irreversible intestinal failure manifesting as severe intractable diarrhea. Thus, most of the patients have a lifelong dependency on PN. As a result, early small bowel transplantation has been suggested for patients with TE ^(7,8). Of 7 patients with TE managed at our institution, 3 were definitively weaned off PN. Follow-ups in these 3 patients range from 4 to 12 years at the time of writing.

Few longitudinal data are available on patients with TE ^(10–12). As a result, there is little basis for predicting the clinical outcome. If bowel transplantation is to be introduced as a treatment option for TE, then the ability to predict the outcome is crucial. To date, no criteria for predicting the duration or degree of dependency on PN are available.

In early case series of TE, mortality was high within the first few months of life, despite progress in nutritional management ⁽⁴⁾. In a 1998 review that included 10 patients with TE, 4 of these patients died before 3 years of age ⁽¹⁰⁾. Most patients had severe intestinal malabsorption and diarrhea requiring long-term total PN (TPN) ⁽¹⁰⁾. TPN is associated with complications that may be fatal or result in small bowel transplantation being performed. Definitive weaning off PN seems rare in patients with TE. We are aware of a single reported case, in a pregnant Maltese woman ⁽⁹⁾; however, the phenotype of TE in Maltese individuals seems milder than in other ethnic groups ⁽⁵⁾. In our study, 3 of 7 patients were definitively weaned off PN. Of the 4 remaining patients, 2 are partially and 2 are totally dependent on PN. Thus, the outcome of the disease varied considerably in our patient population. All 7 patients had typical clinical features with neonatal-onset diarrhea that persisted despite TPN, as well as the above-described typical histological features ^(2,10).

When we compared patients, we found no evidence of a correlation between the initial histological findings and the clinical severity of the disease. Severe histological lesions within the first few years of life were not always followed by continued severe disease. Also, some of the patients with focal lesions initially experienced an increase in the severity of their clinical manifestations over time. When we compared biopsies over time, however, an improvement in the histological lesions with a decrease in tufting was seen in the 3 patients who were weaned off PN, although characteristic lesions were still present in all 3 patients at the time of weaning. In 1 of these patients (no. 1), a few histological lesions were still visible 4 years after weaning from PN. In the 4 patients who remained dependent on PN, the epithelial lesions were variable, usually focal initially, and showed little change over time. The number of patients in our series is too small to allow firm conclusions, however.

Our data suggest that phenotypes and outcomes may vary substantially across patients with TE. In keeping with this possibility, there have been several reports of TE in patients with birth defects such as choanal atresia, anal defects, or skeletal defects ^(13–15). Superficial punctate keratitis has been found in a substantial proportion of patients with TE ⁽⁵⁾. Conceivably, these abnormalities, none of which was present in any of our patients, may be associated with worse outcomes. Another possible explanation for the variability of outcomes across patients may lie in differences at the cellular and molecular levels. Abnormalities of extracellular matrix (ECM) components and basement membrane molecules may be involved in the pathophysiology of TE. Basement membrane abnormalities have been described in children with TE ⁽³⁾. Basement membrane molecules may play a key role in the development and maintenance of the intestinal epithelium, which involves continuous cell generation and proliferation from a stem cell population at the bottom of the crypts, differentiation of the new cells as they migrate along the adjacent villi, and, finally, extrusion of the cells at the tip of the villi ^(16–18). Decreased laminin deposition in the chorion near hyperplastic crypts in TE was found in 1 study but not in another ^(13,19). Abnormalities in α2β1 integrin, a key to interactions between laminin and epithelial cells, have also been postulated ⁽¹⁹⁾. Abnormal distribution of a specific integrin type is an interesting hypothesis, especially given a report of α6β4 deficiency in a child with intractable diarrhea ⁽²⁰⁾. Furthermore, abnormal interactions on enterocytes may increase the immunohistochemical expression of desmoglein and produce ultrastructural alterations in desmosomes ⁽³⁾. Sivagnanam et al ⁽⁶⁾ identified a mutation in the EpCAM gene in 2 affected double second cousins of Mexican descent, and found different EpCAM mutations by directly sequencing the gene in 2 unrelated patients with TE and no mutation in a patient with a less severe form of TE. These findings suggest genetic heterogeneity. It seems that various mutations can lead to a phenotype of TE. Thus, clinical and histological expression of TE could correspond to different cellular and genetic mechanisms. To date, no molecular studies have been done in our patients.

In conclusion, patients with TE may experience clinicopathologic improvements, justifying repeated duodenal biopsies to monitor the histological abnormalities. Furthermore, our finding of definitive weaning from PN in 3 of 7 patients indicates a need for the utmost caution when considering early intestinal transplantation in children with TE ^(21–23). We identified no histological differences across patients that predicted weaning off PN, perhaps because of our small sample size. The patients who were weaned off PN showed improvements in their histological lesions over time. Direct DNA sequencing in our patients will allow us to determine whether the 3 weaned patients share the same type of mutation and whether this mutation differs from the findings in the other 4 patients. The identification of new mutations will improve our understanding of TE and may assist in predicting the outcome.