Nodular regenerative hyperplasia in X-linked agammaglobulinemia: an underestimated and severe complication

Abstract

Background:

Late-onset complications in X-linked agammaglobulinemia (XLA) are increasingly recognized. Nodular regenerative hyperplasia (NRH) has been reported in primary immunodeficiency but data in XLA are limited.

Objective:

to describe NRH prevalence, associated features and impact in patients with XLA

Methods:

Medical records of all XLA patients referred to the NIH between 1994 and 2019 were reviewed. Liver biopsies were performed when clinically indicated. Patients were stratified into NRH+ or NRH− groups, according to their NRH biopsy status. Fisher’s exact test and Mann-Whitney test were used for statistical comparisons.

Results:

Records of 21 patients with XLA were reviewed, with a cumulative follow-up of 129 patient-years. Eight patients underwent at least one liver biopsy of whom 6 (29% of NIH XLA cohort) were NRH+. The median age at NRH diagnosis was 20y (17–31). Among patients who had liver biopsies, alkaline phosphatase (ALP) levels were only increased in NRH+ patients (p=0.04). Persistently low platelet count (<100k/μL for more than 6 months), mildly to highly elevated hepatic venous pressure gradient (HVPG) and either hepatomegaly and/or splenomegaly were present in all NRH+ patients. In opposition, persistently low platelet counts were not seen in NRH− patients, and hepatosplenomegaly only observed in one NRH− patient. HVPG was normal in the only NRH− patient tested. All-cause mortality was higher among NRH+ patients (5/6, 83%) than in the rest of the cohort (1/15, 7% among NRH− and Unknown patients, p=0.002).

Conclusion:

NRH is an underreported, frequent and severe complication in XLA, which is associated with increased morbidity and mortality.

Capsule summary

We report a cohort of 21 XLA patients followed at NIH. NRH was a relatively common complication diagnosed during adulthood (29%) which correlated with increased morbidity and mortality factors.

Graphical Abstract

Introduction

In 1952, Dr. Ogden Bruton wrote a single author, single word title manuscript (“Agammaglobulinemia”) describing what is assumed by most of the clinical immunology community to be the first report of a primary immunodeficiency (PID) or inborn error of immunity (IEI) (1). Almost seventy years later, increased awareness along with readily available diagnostic tools led to the identification of hundreds of X-linked agammaglobulinemia (XLA) cases (2).

XLA is caused by loss-of-function mutations in BTK that impair B-cell development in the bone marrow, resulting in near absence of B cells and profoundly low serum immunoglobulin levels (3, 4). The lack of antibodies confers increased susceptibility to infections, mainly due to encapsulated bacteria, but also enteroviruses and mycoplasma, among other microorganisms. Recurrent and severe infections affecting predominantly the respiratory and gastrointestinal tracts are the most common clinical complications (5).

Immunoglobulin replacement has been the main and sometimes the only therapeutic option for these patients since the initial description of the disease in 1952. When infectious complications could be prevented and controlled, this new scenario allowed for longer survival as well as for the recognition of late-onset, non-infectious complications associated with the disease (6, 7).

In a report of 201 U.S. XLA patients from 2006, infectious diseases were the most common initial presentation as well as the most frequently reported cause of death (8). In a more recent Italian XLA cohort study published in 2020, infections were again the most common phenotype at presentation and follow up (9). Interestingly, among the 13 deceased patients, patently infectious or non-infectious causes of death were more evenly distributed. (9).

Liver disease, however, is rarely reported in the literature among patients with XLA. Conversely, in common variable immunodeficiency (CVID), another predominantly antibody deficiency, chronic liver disease, particularly nodular regenerative hyperplasia (NRH), has been identified as a frequent complication associated with increased morbidity and mortality (10–15). In a retrospective study of 23 patients with primary hypogammaglobulinemia and hepatic abnormalities with liver biopsies, NRH was found in 84% of CVID patients and in the only XLA patient biopsied (10).

Nodular regenerative hyperplasia (NRH) is an underappreciated cause of non-cirrhotic portal hypertension. It is characterized by the diffuse transformation of normal hepatic parenchyma into small (<3mm), regenerative nodules with little to no fibrosis between the nodules (16). The pathophysiology underlying NRH remains unclear but several autopsy studies and case reports have reported a strong association between NRH and vascular abnormalities (17).

Clinically, patients with NRH can be asymptomatic until portal hypertension develops, and most often present with portal hypertension-associated complications, including gastroesophageal variceal bleeding, thrombocytopenia, and hepatosplenomegaly. Liver biopsies are necessary to confirm the NRH diagnosis. Management of NRH is directed towards treatment of underlying disease, if established. As many patients present with complications of portal hypertension, the mainstay of management is prevention and treatment of these conditions, namely variceal hemorrhage, as ascites and hepatic encephalopathy are rare due to preserved hepatic synthetic function. The outcome of NRH is dependent on portal hypertension severity, associated systemic disease, and extremely rarely, risk of rupture of the hyperplastic regenerative nodule (16, 18, 19).

This retrospective study describes our experience with NRH, its prevalence, associated features and impact in patients with XLA.

Methods

We retrospectively reviewed the medical records of all XLA patients referred to the Primary Immunodeficiency Clinic at the National Institutes of Health (NIH) Clinical Center between 1994 and 2019. All study participants and/or the parents/tutors of minor patients provided assent and informed consent, respectively, to different NIH research protocols, which were approved by the National Institute of Allergy and Infectious Diseases (NIAID) institutional review board (IRB). Patients were followed per protocol schedule or more frequently if clinically indicated. Medical care was provided on an outpatient and inpatient basis as necessary. Demographic, clinical, laboratory, pathological, and radiological data were recorded. Hepatology evaluation and liver biopsies were performed when clinically indicated. Persistently elevated liver enzymes (5/8), hepato/splenomegaly (4/8) and thrombocytopenia/declining platelets (4/8) presenting alone or in any combination, triggered liver biopsies in these patients. Of the eight liver biopsies performed, 7 were obtained by transjugular route and one was obtained percutaneously. Biopsy specimens were processed by standard methods, including reticulin staining, and reviewed by a single hepatopathologist (DEK); those that met the criteria of 3+ nodularity and 0–1 fibrous septa were classified as nodular regenerative hyperplasia (16, 20). Patients were stratified into NRH positive (NRH+) or NRH negative (NRH−) groups, according to their NRH biopsy status (patients with no liver biopsies were classified as NRH Unknown). Imaging studies (computed tomography [CT], magnetic resonance imaging [MRI], ultrasound [US]) of all patients were reviewed by a radiologist (JM) blinded to clinical data on each patient. Hepatomegaly and splenomegaly were defined based on pediatric or adult reference ranges as described elsewhere(21, 22). The portal vein diameter and flow direction, as well as porto-systemic collateral circulation were noted. A normal portal vein diameter was defined as 1.3cm or less. A normal portal vein velocity was defined as 20–40cm/s. Imaging studies were assessed for liver nodularity, echogenicity, and echotexture. Laboratory values were presented as medians. Repeated laboratory measurements within the same week were averaged. Baseline and end of follow-up analyses were averaged from the first and last 3 months of follow-up, respectively. Fisher’s exact test was used to compare categorical variables and Mann-Whitney test to compare continuous variables. P-values ≤ 0.05 were considered statistically significant. Statistical analyses were conducted using Stata Statistical Software, Release 14 (College Station, TX: StataCorp LP).

Results

Patient cohort

Records of 21 patients with XLA were reviewed. Patients were referred to NIH between October 1994 and June 2019 for one or more of the following reasons: XLA diagnosis confirmation/second opinion/self-referred (n=7); management of uncontrolled infections (n=7); cytopenias of unclear etiology (n=4); management of suspected autoimmune/inflammatory conditions (n=5); and/or severe reactions to IgG replacement therapy (n=1). Liver disease was not an indication for any of the referrals. All patients were male, 15 were Caucasian, 2 Asian, 3 Black, and 1 Hispanic. The median age of symptoms onset was 12.5 months (n=14 patients) with a median age at XLA diagnosis of 24 months (range 0–9 years; n=19 patients). The first visit to the NIH occurred at a median age of 21 years (range 2–53 years). Patients were followed up at the NIH Clinical Center for a median of 5 years (range 0.1–20 years). The cumulative follow-up in this cohort was 129 patient-years. The diagnosis of XLA was genetically confirmed in 20 patients. Sixteen of these BTK mutations are shared by other XLA patients in the literature, and 4 are novel variants ( See Table E1 in the Online Repository). One patient had both BTK and CYBB pathogenic variants. Initial presentation, timing of diagnosis, main clinical manifestations as well as immunoglobulin replacement in this cohort were similar to previously published XLA series (Table I).

Table 1.

Clinical presentation of patients with XLA at first visit to NIH.

	NIH cohort	N of NIH patients	Previously published series*
Presentation and diagnosis
Age at onset of symptoms, median	12.5 mo	14/21	8–26 mo
Age at diagnosis, median	24 mo	19/21	24–84 mo
Infections
Recurrent sinopulmonary infections	18 (86)	19/21	(85–94)
Gastrointestinal	4 (19)	19/21	(19–53)
Skin	4 (19)	19/21	(8–45)
Septic arthritis/osteomyelitis	3 (14)	19/21	(3–42)
Meningoencephalitis	1 (5)	19/21	(5–28)
Noninfectious complications
Chronic lung disease	7 (33)	19/21	(15–67)
Crohn-like colitis	2 (9.5)	19/21	(3–9.5)
Eosinophilic esophagitis	1 (5)	19/21	—

Values are n (%) unless otherwise indicated.

^*See references 2, 8, 9, and 27–35.

Persistent elevation of liver enzymes, thrombocytopenia/declining platelets, and hepato/splenomegaly were the reasons most frequently mentioned in patients charts for pursuing a liver biopsy in this cohort. Eight patients (38%) underwent at least one liver biopsy and were included in the extended analysis (Figure 1). Search for hepatotropic viruses by means of nucleic acid or antigen evaluation resulted negative (except for one positive blood adenovirus test) in blood, liver or stools when determined among the eight biopsied patients (blood: hepatitis B virus, 8/8 negative; hepatitis C virus, 7/7 negative; Epstein-Barr virus, 7/7 negative; cytomegalovirus, 7/7 negative; human herpes virus 8, 1/1 negative; parvovirus B19, 3/3 negative; and adenovirus, 2/3 negative -liver biopsy tested negative in the patient with the positive blood test-; stools: enterovirus, 4/4 negative). The remaining 13 patients (no liver biopsies performed; NRH unknown) were excluded from NRH-specific extended analysis. Among the 8 biopsied patients, three had prolonged follow-up gaps (lasting more than 5 years) immediately following their initial visit. These gaps were excluded from the analysis.

Figure 1.

Timeline of liver biopsies in the NIH XLA cohort.

Dashed lines represent duration of follow-up not included in the analysis because of unknown NRH status. NRH = nodular regenerative hyperplasia; Pt= patient.

NRH diagnosis

Six patients were diagnosed with NRH, representing 29% of the total NIH XLA cohort. None of the NRH+ patients had been exposed to medications classically associated with NRH (16). The median age at NRH diagnosis was 20 years (17–31). NRH diagnosis was established in the first liver biopsy in all but one patient. A second confirmatory biopsy was needed in one patient due to insufficient material in the first sample. NRH+ patients had a median follow-up of 4 years (9 months – 13 years). The two patients who were negative for NRH had one liver biopsy each, performed at the ages of 23 and 49 years, respectively. The median follow-up of NRH− patients was 4 years (1 – 7 years). Despite being followed for longer periods, the last liver biopsy in the NRH− group was considered as their end of follow-up to ensure their NRH− status when performing comparisons between groups (Figure 1).

Histology

Among the eight patients who had liver biopsies, none of the specimens showed signs of direct viral cytopathic effects or virally triggered immune responses. No acid-fast bacilli were identified on any specimen either. No piecemeal necrosis or significant portal fibrosis was observed in any patient. Liver architectural abnormalities included macrovesicular steatosis in one patient who was on doxycycline therapy; another patient had borderline steatohepatitis with associated mild periportal and perivenular fibrosis however was also on prednisone therapy at the time for treatment of granulomatous skin reaction. Five out of the six patients with NRH also had perisinusoidal fibrosis and perisinusoidal dilatation (likely representing advanced liver disease), and portal hypertension (hepatic venous pressure gradient [HVPG] >5mmHg) that was identified in all patients with NRH. No central venopathy was found in any patient, though portal venopathy was present on liver biopsy of one patient who was also found to have angiosarcoma (Patient 3, see Table E2 in the Online Repository). Two patients with NRH had evidence of early bridging fibrosis as well as chronic cholestasis, one of whom also had evidence of peri-portal copper accumulation. (see Table E2 in the Online Repository). Altogether, the histopathologic presence of fibrosis (perisinusoidal, periportal), cholestasis, perisinusoidal dilatation and portal venopathy were strongly associated with NRH (positive predictive values and specificity, 100%). Classical histopathological findings of NRH in a patient with XLA are depicted in Figure 2.

Figure 2.

Nodular regenerative hyperplasia in a patient with X-linked agammaglobulinemia.

A. There is vague nodularity with areas of sinusoidal dilation and mild, patchy, sinusoidal lymphocytic infiltrates (H&E, 100x). B. The Masson stain shows delicate perisinusoidal fibrosis in areas of sinusoidal dilation (Masson Trichrome, 400x). C. The reticulin framework is distorted by nodular areas of wide liver cell plates bounded by narrowed plates. (Reticulin, 200x). D. A biopsy taken 7 years later shows persistence of nodularity. (Reticulin, 200x).

Liver biochemistry tests

Alanine aminotransferase (ALT) levels were mildly elevated in both NRH+ and NRH− groups, while alkaline phosphatase (ALP) levels were only increased in NRH+ patients (p=0.04) (Table 2). ALP elevations in this group were persistent (lasting more than 6 months) and ranged in severity: mild (up to 2x the upper limit of normal [ULN]) in 2, moderate (2–5x ULN) in 2, and severe (more than 5x ULN) in 2. Both NRH+ and NRH− groups had similar aspartate aminotransferase (AST) levels at baseline but higher values were observed at the end of follow-up in the NRH+ group (85 vs. 32 U/L, p=0.04) (Table 2). Median gamma-glutamyl transferase (GGT) levels were above normal range in NRH+ patients at baseline (97 U/L) and persisted elevated at the end of follow-up (292 U/L, data available from 4/6 patients at both observation points). In the NRH− group, serum GGT level was available from one patient at baseline only and was within normal limits (52 U/L). Total bilirubin levels were normal at baseline among both NRH+ and NRH− patients (Table 2). At the end of follow-up, the NRH+ group had increased median levels of total bilirubin (2.7mg/dL) driven by direct bilirubin which was elevated in 4/6 patients. An elevated direct bilirubin can result from a variety of processes including infection, sepsis, medications, or biliary tree obstruction. The increase in bilirubin seen in these patients was observed over the course of months preceding these patients’ deaths, while they were very sick. Therefore, these rises in bilirubin are likely multifactorial.

Table 2.

Laboratory evaluation at baseline and end of follow-up, according to NRH status.

BASELINE
Hemoglobin level, (13.7–17.5 g/dL)	10.8 (8.9–13.1)	12.8 (12.5–13)	0.4
White blood cell count, (4.230–9.070 K/μL)	3.234 (1.897– 4.400)	4.191 (3.168–5.213)	0.3
Platelet count, (161–347 K/μL)	65 (36–136)	226 (212–240)	0.04
Prothrombin time, (11.6–15.2 s)	14.5 (13.5–15.1)	13 (13–13.4)	0.04
ALT level, (0 – 41 U/L)	49 (22–93)	45 (15–75)	0.7
AST level, (0 – 40 U/L)	54 (37–175)	39 (23–54)	0.3
ALP level, (40–130 U/L)	168 (141–259)	74 (72–75)	0.04
Gamma-glutamyl transferase level, (0–59 U/L)*	97 (45–211)	52	0.5
Total bilirubin level, (0.0–1.2 mg/dL)	0.7 (0.4–0.9)	0.6 (0.5–0.6)	0.5
Albumin level (3.7–4.7 g/dL)	3.4 (2.6–3.8)	3.8 (3.7–3.8)	0.1
Creatinine level, mg/dL (0.67 – 1.17)	0.7 (0.4–1.7)	0.9 (0.7–1.1)	0.4
Blood urea nitrogen level, (8–22 mg/dl)	12 (9–29)	13 (10–16)	0.8
END OF FOLLOW-UP
Hemoglobin level, (13.7–17.5 g/dL)	9.1 (7.6–12)	13.6 (12.1–15)	0.04
White blood cell count, (4.230–9.070 K/μL)	3.550 (0.605–8.196)	4.568 (3.335–5.800)	0.7
Platelet count, (161–347 K/μL)	64 (36–73)	160 (152–168)	0.04
Prothrombin time, (11.6–15.2 s)	14.5 (14–15)	13	0.03
ALT level, (0 – 41 U/L)	44 (39–123)	47 (39–54)	0.6
AST level, (0 – 40 U/L)	85 (48–253)	32 (31–32)	0.04
ALP level, (40–130 U/L)	304 (160–874)	65 (61–68)	0.04
Gamma-glutamyl transferase level, (0–59 U/L)*	292 (110–583)	NA
Total bilirubin level, (0.0–1.2 mg/dL)	2.7 (0.7–21.1)	0.6 (0.5–0.7)	0.06
Albumin level (3.7–4.7 g/dL)	3 (2.1–4.3)	3.9 (3.7–4.1)	0.2
Creatinine level, mg/dL (0.67 – 1.17)	0.7 (0.4–0.7)	0.9 (0.8–1)	0.03
Blood urea nitrogen level, mg/dL (8–22 mg/dl)	20 (10–48)	14 (12–15)	0.3

^*Gamma-glutamyl transferase levels available from 4/6 NRH+ patients at both observation points and 1/2 NRH− patients at baseline.

ALP= alkaline phosphatase; ALT= alanine aminotransferase; AST= aspartate aminotransferase; NA= not available; NRH= nodular regenerative hyperplasia. In italics, abnormal and statistically significant results.

Platelet count

Reduced platelet counts were observed at baseline and at the end of follow-up in the NRH+ group. In both observation-points median values were significantly lower compared to the NRH− group (65 vs. 226 K/μL at baseline and 64 vs. 160 at the end of follow-up, p=0.04 in both comparisons) (Table 2). Longitudinal evaluation of platelet counts revealed a unique pattern of persistently low platelets (<100 K/μL for more than 6 months) among NRH+ but not NRH− patients (Figure 3). Persistence of the above-mentioned platelet phenotype could not be assessed in patient 5 (NRH+) because laboratory data on this patient were only available for 2 months after a downtrend in platelets was noticed.

Figure 3.

Longitudinal evaluation of platelet counts according to NRH status.

Y axis represents platelet counts (K/uL); X axis represents years of follow-up. Inverted triangles correspond to liver biopsies. NRH = nodular regenerative hyperplasia.

Endoscopic and Imaging findings

Esophagogastroduodenoscopy (EGD) studies were available in 8/21 patients. Of the NRH+ patients, 3/6 (50%) had esophageal varices (grade 1). All six NRH+ patients had mild portal hypertensive gastropathy. HVPG was elevated in all NRH+ patients, and normal in the only NRH− patient tested. Of the two NRH− patients, one patient had mild portal hypertensive gastropathy but HVPG was not obtained.

Imaging studies were available for 20/21 patients, including CT, MRI, and US. In the NRH+ group, hepatomegaly and/or splenomegaly were found in all patients. An increased portal vein diameter and collateral circulation (umbilical vein recanalization, esophageal or gastric varices) were found in 5/6 NRH+ patients. Increased echogenicity was found in 3/6 NRH+ patients, whereas increased echotexture was found in 4/6 patients. No liver nodularity was visible in any NRH+ patients. Of the two NRH− patients, one had both hepatomegaly, splenomegaly, and increased portal vein diameter (the same patient who had portal hypertensive gastropathy), however neither had abnormal echogenicity, echotexture, or liver nodularity ( see Table E3 in the Online Repository). Altogether, imaging and/or endoscopic confirmation of hepatomegaly, splenomegaly, collateral circulation, increased liver echogenicity and/or increased echotexture were strongly associated with NRH (positive predictive values and specificity, 100%); while absence of hepatomegaly and/or splenomegaly had a negative predictive value of 100%.

Clinical Outcome

All-cause mortality was higher among NRH+ patients (5/6, 83%) than in the rest of the cohort (1/15, 7% among NRH− and Unknown patients, p=0.002). Intracerebral hemorrhage, upper GI bleed from gastric ulcer, and disseminated intravascular coagulation were the causes of death for 3/6 NRH+ patients. T cell disorders including T cell leukemia and malignant T cell lymphoproliferative disorder in the liver developed in 2/6 NRH+ patients, both of whom died. Of the NRH− patients, one patient died from progressive neurological decline from presumed viral encephalitis; the other remains alive. The NRH+ patient who remains alive has evidence of compensated portal hypertension and is currently undergoing evaluation for liver transplant (Table 3).

Table 3.

Clinical Outcome of NRH+ and NRH− patients.

	Patient number	Clinical course	Cause of death	Age at the time of death
NRH+	1	Development of malignant T-cell lymphoproliferative disorder with kidney, liver, and bone marrow involvement	Severe upper GI bleed with multiple gastric ulcers, sepsis, multi-organ failure	28
	2	Compensated portal hypertension	N/A	Alive
	3	Metastatic angiosarcoma, multifocal aspergillosis affecting lung and brain	Resection of neuroaspergillosis complicated by intraparenchymal/subdural hemorrhage with unsuccessful craniectomy	30
	4	Large granular T cell leukemia	Aspiration pneumonia, progressive multifocal leukoencephalopathy	24
	5	Adenovirus viremia, multiple pulmonary infections	Multi-organ failure, disseminated intravascular coagulation	31
	6	Recurrent pulmonary infections, sinusitis	Enterococcal bacteremia, cerebral aspergillosis, and fungal encephalitis with multiple brain infarcts leading to disseminated intravascular coagulation	29
NRH−	7	Chronic encephalitis of presumed viral etiology	Progressive cognitive decline	29
	8	Recurrent Campylobacter jejuni infections	N/A	Alive

N/A= not available; NRH= nodular regenerative hyperplasia

Discussion

X-linked agammaglobulinemia was one of the first PID/IEI genetically identified, with hundreds of patients reported (and likely even more diagnosed and followed) to date. Originally, uncontrolled infections would lead to early mortality, but timely diagnosis, improved antimicrobial therapies and immunoglobulin replacement currently allow long-term survival into adulthood. As a consequence of this longer lifespan, later-onset complications of XLA are being seen. Despite the high number of XLA cases reported worldwide, data on non-infectious complications of XLA are few as are the reported cases of XLA NRH (10, 24). Chronic liver dysfunction is a well-documented complication in other primary PID/IEI such as CVID (12, 25).

We identified NRH in 6/21 (29%) patients with XLA followed at the NIH. The prevalence of NRH in this cohort supports the association between NRH and immunological diseases (25, 26). Despite the risk of ascertainment or referral bias, our cohort of patients at NIH is similar to XLA cohorts followed elsewhere (2, 8, 9, 27–35). The age at symptoms onset, the age at diagnosis, the underlying BTK mutations, the type and frequency of complications experienced are not exclusive of the NIH cohort or patients with XLA and NRH. Moreover, liver disease was not an indication for these referrals to NIH.

Liver biopsies in this retrospectively analyzed cohort were performed based on clinical indications and showed that NRH was common in XLA patients with liver abnormalities. Considering that NRH can remain asymptomatic for prolonged periods of time (16), it is alarming that the true prevalence of this complication in the full cohort of patients with XLA might be higher than previously estimated.

Published liver histology data in patients with XLA is limited. In a study of patients with predominantly antibody deficiencies presenting with hepatic abnormalities, Malamut et al., reported NRH in the only XLA patient who underwent a liver biopsy (10). Another series, reported by Pulvirenti et al., found increased portal vein diameter in 2 out of 6 patients with XLA. One of the two underwent a liver biopsy that confirmed non cirrhotic portal hypertension with histology suggestive of NRH (24). However, large published series of XLA patients have not examined NRH as a frequent complication of the disease. Among 200 XLA cases registered in the United States Immunodeficiency Network (USIDNET), NRH has been reported in only one patient (30). In a recent survey conducted by the World Allergy Organization, including data from 783 patients with XLA from 40 centers around the world, neither NRH nor liver disease were part of the structured questionnaire of the survey (although an open field for reporting of unusual manifestations was provided) (2). Similarly, in CVID a much higher prevalence of NRH has been reported when detailed evaluation of liver involvement was carried out (11–13).

Several factors might contribute to such contrasting statistics. Nodular regenerative hyperplasia commonly remains asymptomatic until portal hypertension is established (16). The stealthy nature of the disease often causes its diagnosis to be delayed or missed. The necessity of an invasive procedure as well as a knowledgeable pathologist to establish the diagnosis of NRH are critical barriers (36). Furthermore, NRH has been found retrospectively at autopsy study of patients with chronic disease who suffered from infectious complications near the time of death (37). The prevalence of NRH in PID/IEI, and its strong association with infection and death, suggest that the development of liver disease, such as NRH, may provide a secondary “hit” to the immune system in individuals with PID/IEI, thus predisposing them to infection and death from non-liver etiologies.

In our cohort, all NRH+ patients were symptomatic and had histological changes compatible with NRH by the time of their first liver biopsy. Because of the retrospective nature of this study, we could not identify the change of an NRH− to NRH+ liver biopsy. Even for established diagnoses, no single noninvasive procedure attempted was able to accurately distinguish NRH+ from NRH− patients. Liver biochemistry tests, when abnormal, were in the mild to moderate range but attributed to ongoing infections or medications. Hepatosplenomegaly, on the other hand, is not expected in XLA and was surprisingly frequent. Though not specific, splenomegaly and thrombocytopenia are features of portal hypertension and NRH should be included in the differential diagnosis of these manifestations. Portal hypertension is a late development in the process and although following these clues might be helpful in the context of undiagnosed patients, optimal interventions should happen earlier. While it is unclear whether portal hypertension can be prevented, early identification allows for more proactive management including regular endoscopy and bleeding avoidance that ultimately could improve patient’s quality of life and survival. Thus, for clinicians taking care of XLA patients, persistently elevated liver enzymes, presence of hepatomegaly, splenomegaly and declining or low platelets should rise the suspicion of NRH. In this setting and based on our center’s experience and awareness for NRH in XLA and other PID/IEI (12, 38, 39, 42) a diagnostic algorithm is suggested in Figure 4.

Figure 4.

Suggested diagnostic algorithm for NRH in XLA.

NRH = nodular regenerative hyperplasia; XLA = X-linked agammaglobulinemia.

While the pathophysiology of NRH remains elusive, its increased frequency in patients with PID/IEI provides hints about mechanisms involved or redundant. The conditions in which NRH is encountered suggest that the integrity of several pathways is relevant to prevent its evolution, including the NADPH-oxidase pathway (chronic granulomatous disease) (38–40), the PIK3CD pathway (activated phosphoinositide 3–kinase δ syndrome [APDS1]) (41), specific antibody responses (CVID)(12), DNA repair mechanisms (ataxia telangiectasia)(42), or BTK pathway (XLA). On the other hand, whereas infectious agents, particularly viruses, have been only sporadically found and rarely associated with NRH pathophysiology (10, 19, 20, 24), this hypothesis remains valid in immunocompromised patients. In this setting, search for rare (e.g., aichivirus) or yet to be described agents should be considered (43, 44) A possible destructive autoimmune process underlying NRH has been suggested by Malamut et al. who found increased CD8+ cytotoxic T cell infiltrates in liver specimens from patients with CVID and NRH (10). Furthermore, Fuss et al. demonstrated that these CD8+ T cells were a source of increased interferon gamma and could underlay this process in CVID (12). It is possible that other factors such as lack or regulatory B cells and consequent T cell autoimmunity may be involved in XLA NRH, but further studies are warranted in order to address these issues. Of note, NRH has also been reported to be associated with non-immunologic diseases/pathways as cystinosis, sickle cell and cystic fibrosis, impacting on both the morbidity and mortality of the affected patients (45–47).

Considering XLA patients’ NRH prevalence, severity, lack of specific treatment and poor outcome, led us to readdress both our diagnostic and treatment approaches in this population. Moreover, hematologic malignancies and other life-threatening non-infectious complications in adults with XLA suggest that immune reconstitution by hematopoietic stem cell transplantation (HSCT) should be explored in greater depth as a curative option(48, 49). Patients with CVID and NRH who underwent isolated liver transplants, had NRH recurrence in the implanted organ (50); on the other hand, HSCT in XLA patients has already been proved successful (51). While the effect immune reconstitution by HSCT could have on preventing or reverting NRH in XLA patients (or other PID/IEI) has yet to be established, NRH prevalence is certainly lower in immunocompetent individuals when compared to patients with PID/IEI. NRH is prevalent in XLA and contributes to poor outcomes; efforts to prevent and preempt it should be explored.

Clinical implications.

Patients with XLA are at increased risk of developing NRH which is both underdiagnosed and underreported. Presence of hepatomegaly, splenomegaly and declining or low platelets should rise the suspicion of NRH.

Abbreviations

ALP

alkaline phosphatase

ALT

alanine aminotransferase

APDS1

activated phosphoinositide 3–kinase δ syndrome

AST

aspartate aminotransferase

CT

computed tomography

CVID

common variable immunodeficiency

EGD

sophagogastroduodenoscopy

GGT

gamma-glutamyl transferase

HSCT

hematopoietic stem cell transplantation

HVPG

hepatic venous pressure gradient

IEI

inborn error of immunity

MRI

magnetic resonance imaging

NIH

National Institutes of Health

NRH

nodular regenerative hyperplasia

PID

primary immunodeficiency

ULN

upper limit of normal

US

ultrasound

USIDNET

United States Immunodeficiency Network

XLA

X-linked agammaglobulinemia