NOD2-Associated pediatric granulomatous arthritis, an expanding phenotype: Study of an international registry and a national cohort in spain

Abstract

Objective

To study the phenotype characteristics of the largest to date cohort of patients with pediatric granulomatous arthritis (PGA) and documented mutations in the NOD2 gene.

Methods

We analyzed merged data from 2 prospective cohorts of PGA patients, the International PGA Registry and a Spanish cohort. A systematic review of the medical records of interest was performed to identify phenotype characteristics.

Results

Forty-five patients with PGA (23 sporadic cases and 22 from familial pedigrees) and documented NOD2 mutations were identified and formed the basis of the study. Of these 45 patients, 18 had the R334W-encoding mutation, 18 had R334Q, 4 had E383K, 3 had R587C, 1 had C495Y, and 1 had W490L. The majority of patients manifested the typical triad of dermatitis, uveitis, and arthritis. In contrast, in 13 patients, the following “atypical” manifestations were noted: fever, sialadenitis, lymphadenopathy, erythema nodosum, leukocytoclastic vasculitis, transient neuropathy, granulomatous glomerular and interstitial nephritis, interstitial lung disease, arterial hypertension, hypertrophic cardiomyopathy, pericarditis, pulmonary embolism, hepatic granulomatous infiltration, splenic involvement, and chronic renal failure. In addition, 4 individuals who were asymptomatic carriers of a disease-causing mutation were documented.

Conclusion

NOD2-associated PGA can be a multisystem disorder with significant visceral involvement. Treating physicians should be aware of the systemic nature of this condition, since some of these manifestations may entail long-term morbidity.

Pediatric granulomatous arthritis (PGA) encompasses 2 diseases with an identical phenotype, the autosomal-dominant Blau syndrome and early-onset sarcoidosis, a sporadic condition. The finding of identical mutations on exon 4 of the NOD2 gene supports the notion that these diseases are actually the same (1–3). Blau (4) described the syndrome that bears his name as a triad of granulomatous boggy polyarthritis, uveitis, and papuloerythematous rash, whereas Jabs et al (5) described a polyarthritis with uveitis, without rash, but with cranial neuropathy. Early-onset sarcoidosis has been known since the 1970s, and like Blau syndrome, it is characterized by granulomatous boggy synovitis, uveitis, and rash, which is indistinguishable from Blau syndrome, albeit without a familial component (6).

Before the discovery of NOD2 gene mutations, there were sporadic reports of organ involvement beyond the typical triad in both Blau syndrome and early-onset sarcoidosis (7–14), thus challenging the initial notion of a restricted phenotype (15). More recently, as a result of our systematic analysis of the phenotype, we and other investigators (16, 17) have been able to confirm most of the previously described, as well as some previously unknown, clinical features among individuals with confirmed NOD2 mutations (18, 19).

In the present study, the data from the International PGA Registry (16) were merged with the data from a previously described Spanish cohort (17) and then analyzed for phenotype characteristics. We describe herein additional clinical manifestations of PGA in a population of 45 patients identified in that analysis.

PATIENTS AND METHODS

Study subjects.

Patients with atypical manifestations of PGA were identified from the International PGA Registry and DNA repository, which was established in 2005 and spans all pediatric granulomatous diseases, and from the Spanish cohort attending the Systemic Autoinflammatory Diseases Unit, Hospital Clinic, Barcelona, which was established through a genetic testing referral system. Contributors who cared for patients of interest were asked to submit to the principal investigator (CDR) de-identified detailed narratives of their medical histories to capture details that were not included in the database.

The inclusion criteria for the International PGA Registry required the presence of granulomatous inflammation in tissue biopsy samples and either arthritis, uveitis, or rash. The individuals from the Spanish cohort were selected from the genetic testing referral system database if they had “Blau mutations” in the NOD2 gene.

Extraction of genomic DNA.

For the International PGA Registry cohort, genomic DNA was obtained directly from the collaborating sites or was extracted from blood samples (or in 1 case, from a buccal swab sample). Blood samples were processed using standard salting-out elution techniques. The buccal swab was processed using a G1N10 Gene Elute Mammalian Genomic DNA Miniprep kit (Sigma-Aldrich, St. Louis, MO) according to the manufacturer's instructions, with the addition of an increased total lysing volume (360 μl) which included an increased proteinase K volume (30 μl) in order to completely cover the buccal swab, and with a decreased elution volume (60 μl) to obtain more concentrated DNA. For the Spanish cohort, genomic DNA samples were obtained from whole blood samples using the QIAamp DNA Blood Mini kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions.

Analysis of NOD2 gene mutations.

All 12 exons and intronic flanking sequences of the NOD2 gene were amplified by polymerase chain reaction (PCR). The PCR products from the International PGA Registry cohort were treated with ExoSAP-IT (USB, Cleveland, OH), and those from the Spanish cohort were purified using a QIAquick PCR Purification kit (Qiagen). Bidirectional fluorescence sequencing was performed using an ABI BigDye Terminators v3.1 Cycle Sequencing kit and an ABI 3100 automatic sequencer (both from Applied Biosystems, Foster City, CA). In addition to direct sequencing, denaturing high-performance liquid chromatography with a WAVE instrument (Transgenomic, Omaha, NE) was used to screen samples from the International PGA Registry for polymorphisms in exons 1, 2, 8, 9, and 12.

Statistical analysis.

For comparisons between patients with wild-type NOD2 and those with NOD2 mutations and between patients with mutations with and without an expanded phenotype, nonparametric statistics (Mann-Whitney U test) and the chi-square or Fisher's exact test were used as appropriate.

RESULTS

A total of 128 patients have been entered to the International PGA Registry and 12 patients have been entered into the Spanish PGA cohort since their inception. A total of 45 affected individuals (22 from familial pedigrees and 23 sporadic cases) with a documented NOD2 mutation were identified from these 2 cohorts and forms the basis of the present study. In addition, we identified 26 patients who had a granulomatous inflammatory disease with wild-type NOD2. A further 4 asymptomatic individuals with a NOD2 mutation were documented. The remaining individuals were asymptomatic relatives of affected patients or the samples are pending.

Patients with a NOD2 mutation.

In this series, the largest to date, 18 patients carried the mutation encoding R334W, 18 carried R334Q, 4 carried E383K, 3 carried R587C, 1 carried C495Y, and 1 carried W490L (Table 1). Of these 45 patients, 21 were female and 24 were male. The average age at onset of PGA was 30 months (range 3–216 months), and the age at recruitment ranged from 4 years to 49 years in the International Registry cohort and from 4 years to 48 years in the Spanish cohort.

Table 1. Demographic, genetic, and clinical characteristics of all symptomatic patients with NOD2 mutations in the combined International Pediatric Granulomatous Arthritis Registry and the Spanish cohort

Patient	Family	Family member*	Sex	Age at onset, months	NOD2 mutation	Clinical features
						Arthritis†	Uveitis‡	Rash	Extended phenotype§
* Sporadic indicates a sporadic case. PGM = paternal grandmother; PU = paternal uncle; PA = paternal aunt. † All patients with arthritis exhibited polyarthritis. ‡ Uveitis was classified as mild (no sequelae), moderate (presence of complications and preserved vision with correction), or severe (significant loss of vision). § Thirteen patients (5, 7, 9, 10, 23, 27, 34, 35, 36, 37, 38, 40, and 43) had an extended phenotype. EN = erythema nodosum; AH = arterial hypertension; GN = glomerulonephritis; LCV = leukocytoclastic vasculitis; HSM = hepatosplenomegaly; ION = inflammatory optic neuropathy; CRF = chronic renal failure. ¶ Four additional family members with mutations but without symptoms have been identified (see Results section).
1	1	PGM	F	–	R334W	Yes	Severe	Yes	None
2	1	Proband	F	5	R334W	Yes	Severe	Yes	None
3	1	Father	M	60	R334W	Yes	Severe	Yes	None
4	1	PU	M	–	R334W	Yes	Severe	Yes	None
5	2	Mother	F	36	R334Q	Yes	Severe	Yes	Liver
6	2	Proband	M	12	R334Q	Yes	Moderate	Yes	None
7	3	Proband	M	22	R334Q	Yes	Moderate	Yes	Liver, EN, fever, lung, sialadenitis, lymphadenopathy, AH
8	3	Sibling 1	M	21	R334Q	Yes	Mild	Yes	None
9	3	Sibling 2	M	21	R334Q	Yes	Mild	Yes	AH
10	3	Mother	F	168	R334Q	Yes	Mild	Yes	Pulmonary embolism
11	4	Proband	M	12	R334W	Yes	Severe	Yes	None
12	4	Mother	F	24	R334W	Yes	Severe	Yes	None
13	5	Sporadic	M	6	R334W	Yes	Mild	Yes	None
14	6	Sporadic	M	50	R334W	No	Moderate	Yes	None
15	7	Sporadic	M	36	R334Q	Yes	Moderate	Yes	None
16	8	Proband	M	12	R334W	Yes	Severe	Yes	None
17	8	Sibling	M	11	R334W	Yes	No	–	None
18	8	Mother	F	24	R334W	Yes	–	No	None
19	9	Sporadic	M	4	R334W	Yes	Severe	Yes	None
20	10	Sporadic	F	12	R334Q	Yes	Moderate	Yes	None
21	11	Sporadic	M	24	R334W	Yes	No	Yes	None
22	12	Sporadic	M	3	E383K	Yes	Severe	Yes	None
23	13	Sporadic	M	6	R334W	Yes	Moderate	Yes	AH
24	14	Sporadic	M	48	E383K	Yes	No	Yes	None
25	15	Sporadic	F	18	W490L	Yes	Mild	Yes	None
26	16	Sporadic	F	11	R334W	Yes	Moderate	Yes	None
27	17	Sporadic	F	72	R334Q	Yes	Moderate	Yes	Pericarditis, GN, fever, LCV, HSM
28	18	Sporadic	M	9	R334W	Yes	Moderate	Yes	None
29	19	Sporadic	M	24	R334Q	Yes	Moderate	Yes	None
30	20	Sporadic	M	36	R334Q	Yes	Mild	Yes	None
31	21	Sporadic	F	4	R334Q	Yes	Mild	Yes	None
32	22¶	Proband	F	48	E383K	Yes	No	Yes	None
33	22¶	PA	F	–	E383K	Yes	Yes	Yes	None
34	23	Sporadic	M	4	C495Y	Yes	Moderate	Yes	Fever
35	24	Proband	F	48	R587C	Yes	Severe	No	Fever, ION
36	24	Sibling	F	21	R587C	Yes	Moderate	Yes	Fever, EN, pericarditis
37	24	Mother	F	216	R587C	No	No	Yes	Fever, EN
38	25	Sporadic	F	11	R334Q	Yes	Severe	Yes	Cranial neuropathy
39	26	Sporadic	F	24	R334Q	Yes	No	Yes	None
40	27	Sporadic	M	22	R334Q	Yes	Moderate	Yes	AH, hypertrophic cardiomyopathy, HSM, CRF
41	28	Sporadic	M	9	R334Q	Yes	No	Yes	None
42	29	Proband	M	20	R334Q	Yes	No	Yes	None
43	29	Mother	F	18	R334Q	Yes	Moderate	Yes	CRF
44	30	Sporadic	F	9	R334W	Yes	No	Yes	None
45	31	Sporadic	F	12	R334W	Yes	No	Yes	None

The large majority of patients (34 of 45) exhibited the complete classic triad of rash, uveitis, and arthritis as originally described for the disease (4). We found 1 patient in the Spanish series and 1 in the International Registry without arthritis, 2 without rash, and 10 without documented uveitis at the time of inclusion. In all but 2 patients from the International Registry, epithelioid granulomas were documented on examination of biopsy tissues (skin and synovium samples, respectively). All 4 Spanish patients with available biopsy tissue showed granulomatous inflammation. In none of the 45 patients were antinuclear antibodies or rheumatoid factor detected in their sera.

Demographic and clinical manifestations considered to be part of the classic phenotype, including polyarthritis, uveitis, and ichthyosiform rash are shown in Table 1. Thirteen of 45 patients in this series (29%) presented with at least 1 manifestation either at disease onset or during the course of the disease that can be considered atypical; these are described in detail below. The age at disease onset was not different in the group with an expanded phenotype as compared with the group with only the classic triad. The median disease duration at entry into the cohort was 227 months (range 54–474 months) in the group with an expanded phenotype and 99.5 months (range 20–528 months) in the group with the classic triad (P < 0.05).

Comparison of the relative frequencies of the various NOD2 mutations between patients with and without an expanded phenotype revealed a relative underrepresentation of R334W in the expanded group (P < 0.05). The other mutations showed no between-group differences. Similarly, there were no differences in either the age or the sex distribution or in the prevalence of the clinical components of the classic triad (arthritis, uveitis, dermatitis) between these 2 groups.

Patients without a NOD2 mutation (wild-type).

Twenty-six patients presented with a granulomatous inflammatory disease and carried wild-type NOD2. Their median age at disease onset was 36 months (range 0–132 months), their age at study entry was 11 years (range 2–42 years). The age at disease onset and the age at study entry were not significantly different between patients with wild-type NOD2 and patients with a NOD2 mutation.

Arthritis, eye inflammation, and skin inflammation were seen in 52%, 64%, and 80%, respectively, of the patients wild-type NOD2. Only 2 individuals exhibited the complete classic triad. The prevalence of arthritis was significantly higher in the group with a NOD2 mutation (P < 0.0001), but not the prevalence of uveitis nor cutaneous inflammation. The clinical features in this group was much more variable; hence, the percentage of patients with expanded manifestations was much larger than that in the group with a NOD2 mutation, which showed a more restricted phenotype, as expected (81% versus 29%; P < 0.0001). These manifestations comprised fever, hepatosplenomegaly, and lymphadenopathy, as well as lung, kidney, gastrointestinal, bone, and central nervous system (CNS) involvement. In addition, this group exhibited Mikulicz syndrome and a large variety of skin manifestations, including erythema nodosum, lobular panniculitis, and cutaneous vasculitis. Only 2 patients in this group presented with the tan-colored skin rash. Fever, hepatosplenomegaly, lung, and CNS involvement were significantly more prevalent in the group with wild-type NOD2 than in the group with a NOD2 mutation.

Expanded manifestations in patients with a NOD2 mutation.

Fever.

The presence of fever was not unusual, and it may have been underreported, since it tends to be mild to moderate. A total of 7 individuals experienced fever, including 1 patient with prolonged fever at presentation, whereas the others experienced fever in association with a disease flare. There were no cases of either periodic or hectic fever.

Panniculitis.

Three patients had subcutaneous nodules, and all of them had febrile episodes. A clinical diagnosis of erythema nodosum was made in these 3 patients, although no histologic confirmation was attempted. In 1 of them (patient 7, Table 1), a single febrile episode was observed during a systemic disease flare, whereas in the others, the fever was recurrent.

Granulomatous lymphadenopathy.

One patient (patient 7) presented with prominent cervical lymphadenopathy. Due to concerns of lymphoproliferative disease, a biopsy of a posterior auricular node was obtained. Multiple epithelioid granulomas were identified on a background of an inflammatory infiltrate, with histiocytes and mature lymphocytes. This case has previously been reported (8, 18).

Liver and spleen involvement.

Patient 7 underwent liver biopsy as part of the protocol for monitoring methotrexate toxicity. He was found to have multiple hepatic epithelioid granulomas, with no evidence of parenchymal inflammation or organ dysfunction after 12 years of followup. Another patient (patient 5) developed liver enlargement; a biopsy obtained during elective abdominal surgery showed noncaseating granulomas. Asymptomatic liver and spleen involvement manifested as hepatosplenomegaly was observed in 2 patients (patients 27 and 40), and in both, this was considered an inconsequential physical finding. No histologic analysis of the liver or spleen was available in either of these 2 patients.

Salivary gland involvement.

Intense granulomatous infiltration of the salivary gland was detected serendipitously by means of a posterior cervical lymph node biopsy in patient 7. To date, he has not developed xerostomia or lacrimal gland involvement.

Leukocytoclastic vasculitis.

Recurrent nonspecific maculopapular and urticarial rash developed in 1 patient (patient 27). Because of the persistence of urticarial lesions, a biopsy was performed, which revealed classic leukocytoclastic vasculitis, with no hemorrhage, and prominent nuclear dust. Episodes resolved after institution of anti–tumor necrosis factor therapy. This patient has been described elsewhere (19).

Pneumonitis.

One patient (patient 6) presented with the acute onset of fever and adenopathy. Computed tomography of the chest revealed mild upper mediastinal, but no hilar, adenopathy. Unexpectedly, several small areas of ground-glass opacity were noted in the medial segment of the right middle lobe and bilateral lower lobes of the lung parenchyma. A repeat scan performed a few months later following intensification of the corticosteroid therapy revealed resolution of both abnormalities. This patient has been described previously (18).

Granulomatous glomerulonephritis and interstitial nephritis.

Patient 27 had a 6-year history of classic Blau syndrome when she developed leukocyturia, an increased serum creatinine level, and mild proteinuria (1 gm in 24 hours). A renal biopsy revealed epithelioid granulomas in 3 of 9 glomeruli examined and severe interstitial granulomatous inflammation. Renal function and findings on urinary microscopy normalized after 4 infusions of infliximab (19). Two additional patients (patients 40 and 43) have stable chronic renal insufficiency.

Arterial hypertension.

Two young patients (patients 9 and 23) presented with asymptomatic arterial hypertension soon after the diagnosis of PGA. In both, additional investigations demonstrated normal renal function, normal findings on urinary microscopy, and negative findings on digital angiograms of the renal arteries. Both patients are currently receiving combination therapy that includes β-blockers and angiotensin-converting enzyme (ACE) inhibitors.

Patient 40 already had hypertension at the time PGA was diagnosed (22 months of age). Currently, he has chronic cardiomyopathy and chronic renal failure. One additional patient (patient 7) has recently developed high blood pressure, and like his sibling (patient 9), he was started on ACE inhibitors. Findings of clinical evaluations have been negative, although an angiogram was not performed.

Pericarditis.

Two children (patients 27 and 36) developed acute pericarditis; this condition was suspected based on the findings of chest radiography in 1 of them and was confirmed by echocardiographic findings in both. In these 2 patients, the episode was a single occurrence and was associated with a febrile/systemic flare, which subsided with medical management.

Cranial neuropathy.

Patient 38 had an episode of transient (few weeks) unilateral facial palsy, as well as an inflammatory optic neuropathy secondary to panuveitis.

Pulmonary embolism.

One adult patient (patient 10) with the typical Blau syndrome phenotype developed chest pain. An acute pulmonary embolism of the right pulmonary artery was demonstrated. Investigations to exclude a prothrombotic state or hypercoagulability disorder yielded negative results. This patient had no previous risk factors other than contraceptive use, and after 1 year, during which she has been taking anticoagulants, she has had no recurrence.

Asymptomatic carrier status.

Four members of a family carrying the E383K substitution exhibited no signs or symptoms of disease (data not shown). Although 3 of them are still younger than age 4 years, the fourth is an adult male age 40 years. Two members of this pedigree who were positive for the mutation exhibited the classic phenotype (polyarthritis, uveitis, and rash). A more detailed analysis of this family is reported elsewhere in this issue of Arthritis & Rheumatism (20).

DISCUSSION

This report is the first to provide a detailed description of the expanded phenotype of pediatric granulomatous arthritis exclusively among patients with documented mutations in the NOD2 gene. Our findings endorse the notion that PGA may constitute a systemic granulomatous inflammatory disorder.

In 1985, Jabs et al (5) and Blau (4) described 2 separate multiplex families showing a similar clinical phenotype that included early-onset polyarticular boggy synovitis, severe panuveitis, and an autosomal-dominant pattern of inheritance. The family described by Blau also presented with a tan-colored maculopapular rash, which was absent from Jabs' report, while the family described by Jabs also presented with cranial neuropathy. All patients characteristically showed noncaseating granulomas on tissue biopsy. The Blau syndrome phenotype in another family has also been reported by Pastores et al (15). The initial idea that this disease entailed a restricted phenotype was later challenged by the descriptions of 2 families: one with associated liver granulomatous involvement and the other with renal granulomatous disease (7, 8).

Most investigators currently accept cranial neuropathies, systemic involvement (fever), and arteritis as part of the disease spectrum. Hence, the family described by Jabs et al (5) and those described by Hafner and Vogel (9) and by Rotenstein et al (10) could be considered to fall within the Blau syndrome spectrum. The phenotype of the sporadic form of the disease, formerly known as early-onset (childhood) sarcoidosis, has itself been expanded to include large-vessel vasculitis (11–13) and visceral involvement, including granulomatous infiltration of the lungs, heart, liver, and kidneys (14). Cranial neuropathy had been identified in association with the Blau syndrome before genetic analysis was available (5). We were able to confirm this complication in 1 patient in our series who had a transient facial nerve palsy. The same patient presented with an inflammatory optic neuropathy complicating severe panuveitis.

We observed that leukocytoclastic vasculitis, which presented phenotypically as urticarial vasculitis, is part of the spectrum of PGA. Large-vessel vasculopathy, which was reported by Wang et al (21) to be present in mutation-positive patients and was reported by other investigators before that (11–13), was not observed in our series of patients.

The present study has expanded the list of cutaneous manifestations of the disease. The dermatitis described as a maculopapular rash with ichthyosiform desquamation and dermal granulomas was documented in 40 of the 45 patients; hence, without a doubt, it is a cardinal manifestation. This study confirmed that recurrent erythema nodosum–like panniculitis as well as cutaneous small-vessel vasculitis are also part of the spectrum.

Lymphadenitis (excluding hilar adenitis) and interstitial pneumonitis have not previously been reported in PGA associated with a NOD2 mutation. In addition, sialadenitis, hepatic involvement, and splenic involvement seem to be without clinical consequences, at least in the short-term. These manifestations therefore may not require invasive diagnostic procedures when identified as an isolated physical finding.

Pericarditis has not previously been reported in early disease. In our series, it was symptomatic in 1 patient and was an echocardiographic finding in 1 patient. Cardiomyopathy found in 1 patient could have been the result of chronic hypertension and not a bona fide primary manifestation. Pulmonary embolism found in 1 patient should be regarded with caution, since risk factors for pulmonary embolism can be elusive.

Renal disease was described as interstitial nephritis in a population of patients with childhood sarcoidosis before genetic testing was available (22). Here, we show 1 documented case of acute glomerulonephritis with granulomatous involvement of the glomeruli and interstitium. We observed 2 additional cases of chronic renal failure but without histologic documentation of granulomatous kidney disease.

Hypertension was previously described in association with renal arteritis (10, 12). Our findings confirm that systemic hypertension severe enough to require treatment is not uncommon, even during the early phases (patients 7, 9, and 23) of the disease, and should be monitored closely. The mechanism is unclear, since in the 2 children who were evaluated, digital imaging of the renal vasculature, tests of renal function, and microscopy of the urine yielded normal findings. Still, we cannot rule out silent granulomatous nephritis as the cause of the observed hypertension.

The expanded manifestations were noted both in sporadic and familial cases and seemed to cluster in a subset of patients that constituted more than one-fourth of the cohort. This may suggest the presence of additional phenotype-influencing genes in NOD2-associated PGA, as has been demonstrated for other autoinflammatory diseases (23). In addition, the longer disease duration in patients with an expanded phenotype implies that the classic triad may evolve in a greater proportion of patients with longer followup and that additional features may emerge with time.

In our cohort, we observed some patients who were diagnosed at a later age. Thus, genetic analysis of NOD2 mutations could be considered not only in children with a phenotype suggestive of PGA, but also in adults.

A comparison of clinical manifestations between patients with and without NOD2 mutations revealed that arthritis, uveitis, and skin rash are not seen exclusively in the presence of NOD2 mutations, although the characteristic triad is significantly more common in those with NOD2 mutations. Conversely, although we found extended manifestations in the group with NOD2 mutations, their frequency was significantly higher among individuals with wild-type NOD2. Still, the presence of overlapping clinical features supports the hypothesis that additional genetic factors may contribute to the phenotype in pediatric granulomatous inflammatory diseases.

Our investigation has some shortcomings, most of which are related to the retrospective character of the clinical data collection. First, there may have been an underestimation of the exact frequency of certain subclinical manifestations that were identified as a result of unrelated investigations. Among them, sialadenitis, granulomatous liver disease, and pericarditis should be noted. Second, in the absence of histologic demonstration of granulomas, a causal link between the NOD2 mutation and some of the atypical manifestations, such as arterial hypertension, cardiomyopathy, chronic renal failure, and pulmonary embolism, cannot be proven.

From the strong association of NOD2 with granulomatous inflammatory disease and from the knowledge of its role in innate immunity, one might be tempted to hypothesize that NOD2 plays a pathogenic role in the formation of granulomas. Apart from mutations in the central NACHT domain that are strongly associated with PGA, several mutations and polymorphisms in the N-terminal leucine-rich repeat domain are known to confer susceptibility to Crohn's disease as well. However, as our data demonstrated, granulomatous disease, even in infants, can occur with wild-type NOD2, and mutations can also be silent. The phenotype of individuals with NOD2 mutations is clearly more complex than was initially thought, requiring investigations into potential additional genetic factors capable of modulating the phenotype, although as we have shown here, this clinical variability may be partly related to the duration of the disease.

In conclusion, by analyzing 2 PGA cohorts in combination, we have identified new clinical manifestations of PGA and have confirmed several previously noted manifestations of the disease in patients with a documented NOD2 gene mutation. The obvious systemic nature of this condition may require yet further consideration for an appropriate name, such as juvenile systemic granulomatosis, which may better describe the nature of the disease.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Rosé had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Rosé, Aróstegui, Martin, Scalzi, Yagüe, Rosenbaum, Wouters.

Acquisition of data. Rosé, Aróstegui, Espada, Scalzi, Rosenbaum, Modesto, Arnal, Merino, García-Consuegra, Carballo Silva, Wouters.

Analysis and interpretation of data. Rosé, Aróstegui, Martin, Rosenbaum, Modesto, Arnal, Merino, García-Consuegra, Carballo Silva, Wouters.

Acknowledgements

The authors thank the study patients and their families for their participation, as well as the following individuals for their contributions to coordinating the samples and conducting the genotyping assays: Trudy M. Doyle, Jinnell A. Lewis, and Carrie Austin. The following physicians have graciously referred their patients to the International PGA Registry: Bernard Lauwerys (Belgium), Ellen Nordal (Norway), Gabriele Simonini (Italy), Rebecca Ten Cate (The Netherlands), Francesco Zulian (Italy), Sheila Olivera-Knupp (Brazil), Dorothee Stichweh (USA), Carol Wallace (USA), Anders Fasth (Sweden), Gloria Higgins and Daryl Kurz (USA), Frank Saulsbury (USA), Marco Gattorno (Italy), Henryka Mazur-Zielenska (Poland), and Françoise Berthet (Luxembourg).