Introduction

Celiac disease (CD) is common disorder caused by an immunologically mediated response to gluten in susceptible persons. As a consequence of genetic susceptibility, patients generate cytotoxic T cells directed against gliadin peptides resulting damage to the small intestine [1].

Aberrant expression of tissue transglutaminase (TTG), which mediates the deamidation of gliadin peptides, is an important component of the pathogenesis of CD. Patients with CD have high titres of antibodies to TTG and deamidated gliadin (DGP) peptides. These antibodies are used as serological markers of CD and are helpful in making the diagnosis [2]. The gold standard remains a duodenal biopsy, showing the characteristic histological features including lymphocyte infiltrates, crypt hyperplasia and villus atrophy [1]. The Marsh-Oberhuber grading system is commonly used to classify the severity of the histological lesions [1].

Celiac disease can present in a variety of clinical contexts. Typically, patients present with abdominal symptoms, including discomfort, bloating, diarrhoea or constipation as well as malabsorption. In children, failure to thrive and iron deficiency are common presentations [3]. There is an increased risk of malignancy, especially of the gut. It has become increasingly apparent that some patients manifest extraintestinal complications including epilepsy, ataxia, dementia, osteoporosis, chronic fatigue, liver dysfunction and unexplained infertility [1]. Early identification of celiac disease can significantly ameliorate some of these complications. Furthermore, identification of the disorder in one individual should prompt a search for the disorder in other members of the family.

Here we describe a patient with undiagnosed celiac disease who was identified after she presented with profound asymptomatic hypogammaglobulinemia. Most importantly, her hypogammaglobulinemia completely resolved on a gluten free diet. There was corresponding improvement in her duodenal histology with normalization of her celiac serology. This case emphasises the importance of identifying celiac disease as a treatable cause of secondary hypogammaglobulinemia. Application of the Ameratunga et al. (2013) diagnostic criteria for Common Variable Immunodeficiency Disorder (CVID) can help identify patients presenting with secondary causes of hypogammaglobulinemia.

Case Report

A 32 year old woman consulted her family practitioner for back pain. The source of the back pain was not clear. During the course of investigations, she was noted to have reduced staining of the gamma region on protein electrophoresis. Subsequent studies showed an IgG of 2.9 g/l (7–14). Her IgA was 0.8 g/l (0.8–4.3) and her IgM was 0.2 (0.4–2.3).

On further questioning, she was in excellent health; specifically, there was no history of recurrent upper or lower respiratory tract infections, or any other symptoms attributed to her hypogammaglobulinemia. Systems review indicated she had mild long-standing abdominal discomfort for which she had not sought medical advice.

Investigations for secondary causes of hypogammaglobulinemia were undertaken. Repeat testing showed an IgG of 4.8 g/l. She had a normal albumin level of 35 g/l (nr 35–50). Her stool alpha 1 antitrypsin was also normal 0.4 mg/g (nr <0.5). Given the mild abdominal symptoms, celiac serology was obtained. Her celiac serology was abnormal, with strongly positive TTG IgA levels as well as elevated DGP antibodies (both > 100 IU/ ml, nr <20).

A duodenal biopsy was undertaken, which showed villus atrophy and crypt hyperplasia (Fig. 1). A lymphocyte infiltrate in addition to villus atrophy (Fig. 1, Marsh class 3) were noted. Subsequently she was placed on a gluten free diet.

Fig. 1
figure 1

Gut histology prior to diagnosis. Showing severe villus atrophy associated with a lymphocyte infiltrate. Many plasma cells are noted (staining red)

Full size image

She was blood group O negative and anti A and anti B isohemagglutinins were detected. Vaccine challenge responses were undertaken before the celiac serology was known. There was an excellent response to both T cell independent (Pnemovax®) and T cell dependent (tetanus, diphtheria and Haemophilus influenzae type B, HIB) vaccines (Table 1). Testing showed 4/23 Pneumococcal serotypes above 1.3 μg/ml before and 17/23 serotypes after vaccination. Review of the gut biopsies showed the presence of plasma cells (Fig. 1). Her metabolic parameters including red cell indices, folic acid and vitamin B12 were normal at the time of presentation. She was iron deficient but this did not affect the red cell indices (Table 2).

Table 1 Important immunology laboratory results
Full size table
Table 2 Metabolic parameters. Hb- Hemaglobin, MCV- mean cell volume, MCH- Mean cell Hb
Full size table

Over the coming months her immunoglobulins increased towards the normal range. In the following year she completed a successful pregnancy with a slight decrease in her IgG. Repeat biopsy showed mucosal healing following her gluten free diet (Figs. 2 and 3). She remains symptomatically well with no increase in susceptibility to infections. Over the following 24 months, her IgG normalised (Fig. 4). Her lgG was 7.2 g/l in August 2015.

Fig. 2
figure 2

Marked improvement in histology following gluten free diet

Full size image
Fig. 3
figure 3

Healing of the duodenal mucosa after a gluten free diet. Persistence of lymphocytic infiltrate is noted

Full size image
Fig. 4
figure 4

IgG levels before and after commencing a gluten free diet

Full size image

Discussion

As serology for CD has become more accurate, it has become apparent that the disorder is much more prevalent than previously suspected [3]. Furthermore, it seems likely there has been a genuine increase in the prevalence of CD in the last 50 years. It is also recognised that CD is not confined to persons of European ancestry alone. Recent data has shown CD occurs in several ethnic groups including Indians, Arabs and Africans [4, 5].

As with many disorders, it has become apparent that the phenotypic spectrum of CD is broader than previously suspected. In addition to the well-recognised gastrointestinal symptoms, patients with CD can present with extraintestinal manifestations including a range of neurological complications, osteoporosis, iron deficiency, liver dysfunction, fatigue and unexplained infertility [1]. Reversible severe hypogammaglobulinemia has not been described as one of the complications of CD.

Several studies have shown an increased risk of CD in patients with other disorders such as type 1 diabetes, Down’s syndrome as well as CVID [6]. CVID is the most frequent symptomatic primary immune deficiency disorder in adults. It probably represents a spectrum of genetic defects culminating in late onset antibody failure (LOAF). It is the most important differential diagnosis in adult patients presenting with severe hypogammaglobulinemia.

The previous (1999) European Society for Immune Deficiency (ESID) and Pan American Group for Immune Deficiencies (PAGID) criteria relied on abnormal vaccine responses and exclusion of secondary causes of hypogammaglobulinemia to establish a diagnosis of CVID. We have discussed difficulties with these criteria and have recently published new diagnostic criteria for CVID, which we believe will improve diagnostic precision (Table 3) [7, 8]. The Ameratunga et al. (2013) criteria are useful in excluding secondary causes of hypogammaglobulinemia with more certainty [9, 10]. In order to meet the criteria for CVID, patients must have significant hypogammaglobulinemia without an obvious secondary cause (category A). Patients must be over 4 years as younger children may have monogenic immune defects or transient hypogammaglobulinemia.

Table 3 Ameratunga et al. (2013) diagnostic criteria for CVID [7]
Full size table

Most importantly, patients with CVID must have in vivo evidence of immune system failure (ISF). The majority of patients with CVID will have a predisposition to infections, autoimmunity or inflammatory disorders (category B). There are also a variety of markers in the peripheral blood including reduction or absence of other immunoglobulin isotypes and a reduction in switched memory B cells which may support the diagnosis. There may be absent isohemagglutinins as well as impaired vaccine responses compared to normal controls (category C). IgG3 deficiency is also considered a biomarker for a humoral immune defect. None of these markers are specific but in combination support the diagnosis. In some patients, there are characteristic histological findings (category D) which support the diagnosis (Table 3) [7]. If there are characteristic histological markers, serological evidence (Category C) criteria are not required for a diagnosis of CVID.

The relationship between CVID and CD is complex. Some patients have both disorders. Identifying CVID patients with concomitant CD can be difficult. Villus atrophy is common in CVID and we have not included this finding in our criteria for CVID [11–13]. In CVID, villus atrophy may be due to a variety of causes including Giardia, autoimmune enteritis, other infections, drug toxicity as well as celiac disease. As the majority of patients with CVID are IgA deficient, IgA-based celiac serology is unhelpful. Likewise, IgG-based CD serology may not be helpful given the humoral immune defect in CVID. Furthermore, IVIG preparations contain IgG DGP antibodies, making this a difficult test to use for diagnosis [14]. Tissue typing is only helpful if the patient is negative for HLADQ2 or DQ8, making CD unlikely [15]. A patient’s response to a gluten free diet appears to be the most helpful marker in identifying patients with CVID who also have CD [14, 15]. Healing of the gut mucosa following a gluten free diet confirms the presence of CD in these patients [6].

CVID no longer needs to be a diagnosis of exclusion and application of the Ameratunga et al. (2013) diagnostic criteria may be helpful in excluding secondary causes of hypogammaglobulinemia [9, 10]. Our patient does not meet laboratory criteria (Categories C or D) for CVID. While she has significant hypogammaglobulinemia, she has preserved vaccine responses. She has isohemagglutinins and memory B cells. Importantly, she has plasma cells in her gut biopsy, making CVID less likely (Fig. 1). If such patients do not meet criteria C or D, CVID is unlikely and this should prompt an aggressive search for other explanations for the hypogammaglobulinemia [10].

In her case, celiac serology was included in the diagnostic work up given her mild long-standing gut symptoms. Following the identification of high titres of TTG IgA antibodies and DGP IgG antibodies, celiac disease was confirmed by gut biopsy. A strict gluten free diet resulted in resolution of her mild long-standing gut symptoms and improvement of her gut histology. In addition, there was resolution of her severe hypogammaglobulinemia. Patients with severe hypogammaglobulinemia associated with CD have been previously described, but there was no confirmation that the IgG normalised following a gluten free diet [16, 17]. It would seem more likely these patients had both CVID and CD rather than purely CD-induced reversible hypogammaglobulinemia as in our patient.

ESID registry has recently published revised (2014) criteria for CVID (http://esid.org/ Working-Parties/Registry/Diagnosis-criteria) which are similar to the Ameratunga at al (2013) criteria [18]. Like the Ameratunga et al. (2013) criteria, the revised ESID registry (2014) criteria may also prove useful in identifying patients with secondary hypogammaglobulinemia.

The precise mechanism of the hypogammaglobulinemia associated with celiac disease remains uncertain in our patient. A significant reduction in albumin can sometimes identify patients with gut or renal loss of protein. Her albumin and stool alpha 1 antitrypsin were within normal limits, making protein loss an unlikely explanation for her significant hypogammaglobulinemia. Nutritional deficiency caused by her celiac disease is also unlikely to have caused the hypogammaglobulinemia. She was noted to be mildly iron deficient, without affecting red cell parameters. Iron deficiency alone is unlikely to explain the severe hypogammaglobulinemia (Table 2). Direct suppression of plasma cell function by cytotoxic/ suppressor cells cannot be excluded in this situation. Direct toxicity of gliaden on intestinal epithelial cells has been shown in vitro [19]. We have not found any studies exploring similar effects on B cell lines.

This is the first case of profound asymptomatic hypogammaglobulinemia as the presentation of CD. Asymptomatic hypogammaglobulinemia should now be added to the ever-increasing list of clinical presentations of celiac disease [1]. All patients with unexplained hypogammaglobulinemia should be screened for celiac disease. As shown here and our previous publications [10, 20], application of the Ameratunga et al. (2013) criteria for CVID may be helpful in identifying patients with secondary causes of hypogammaglobulinemia.