Potential conflict of interest: Nothing to report.
Idiopathic noncirrhotic portal hypertension (INCPH) is characterized by an increased portal venous pressure gradient in the absence of a known cause of liver disease and portal vein thrombosis. In contrast to the high prevalence of this disorder in India, INCPH is a rare disease in the Western world. The etiology of INCPH can be divided in five categories: chronic infections, exposure to medication or toxins, thrombophilia, immunological disorders, and genetic disorders. Multifactorial etiology can also be encountered. Chronic abdominal infection is incriminated as the most important etiological factor in Eastern patients and thrombophilia in Western patients. The majority of patients with INCPH initially present with signs or complications of portal hypertension (mainly variceal bleeding and splenomegaly). These patients usually have preserved liver function. Liver function impairment occurs mainly in the context of intercurrent conditions. Patients with INCPH are often clinically and radiologically misdiagnosed as liver cirrhosis, so that a liver biopsy is indispensable to discriminate cirrhosis from INCPH. Histopathological characteristics of INCPH are heterogeneous, demonstrating overlap between several pathological entities (e.g., hepatoportal sclerosis, nodular regenerative hyperplasia, and incomplete septal cirrhosis). Even though hemodynamical changes in INCPH patients are not comparable to those in cirrhotics, prophylaxis and treatment of variceal bleeding are recommended to be similar. Anticoagulation therapy must be considered only in patients who develop portal vein thrombosis. INCPH has been considered a disorder with a relatively benign disease course. However, liver failure, hepatic encephalopathy, and hepatopulmonary syndrome can occur and are considered indications for liver transplantation. (HEPATOLOGY 2011;)
Portal hypertension is a clinical syndrome defined by a portal-caval venous pressure gradient exceeding 5 mm Hg.1 This increase of portal pressure eventually will lead to the development of collateral circulation and splenomegaly. In the Western world, liver cirrhosis is the most frequent cause of portal hypertension. However, in a variety of disorders, portal hypertension develops in the absence of cirrhosis. This condition, referred to as noncirrhotic portal hypertension, is often classified based on the site of obstruction (i.e., prehepatic, intrahepatic, and suprahepatic portal hypertension) (Table 1). Worldwide, the most common cause of noncirrhotic portal hypertension is schistosomiasis.2 In the Western world, chronic liver diseases, such as nonalcoholic steatohepatitis, primary biliary cirrhosis, primary sclerosing cholangitis, and congenital hepatic fibrosis, together with extrahepatic portal vein thrombosis and Budd-Chiari syndrome, are common causes of noncirrhotic portal hypertension.3-5 If all these conditions have been ruled out, the diagnosis of idiopathic noncirrhotic portal hypertension (INCPH) can be made (Table 2).6 The international nomenclature about this condition is ambiguous. In the Indian subcontinent, this condition is known as noncirrhotic portal fibrosis, whereas in Japan and other Asian countries, it is referred to as idiopathic portal hypertension. In the Western world, this condition has been variably termed hepatoportal sclerosis, idiopathic portal hypertension, incomplete septal cirrhosis, and nodular regenerative hyperplasia (NRH). Because all these entities share histopathological characteristics (e.g., obliterative vascular lesions) and clinical profile, it has been suggested that INCPH can be viewed as a distinct single entity with various pathological aspects, rather than different clinicopathological entities. Agreement on uniform nomenclature is an essential requirement for collaborative studies. We, therefore, suggest that in future studies, the term INCPH should be used, as it covers both the clinical and etiological aspects of the disorder. The aim of this review is to provide a critical appraisal of the available scientific literature of this disorder in the Western world. Additionally, differences and similarities between Western and Eastern patients will be discussed.
Table 1. Causes of Noncirrhotic Portal Hypertension
At the end of the 19th century, Banti described a syndrome characterized by marked splenomegaly and anemia in the absence of hematological disease.7 In retrospect, it becomes clear that the patient cohort studied by Banti comprised patients with cirrhosis, INCPH, and tropical splenomegaly syndrome caused by chronic malaria. Subsequently, a panel of Indian experts denominated splenomegaly in patients without liver pathology or chronic malaria as noncirrhotic portal fibrosis.8, 9 In India, INCPH incidence estimates as high as 23% have been reported.10, 11 In the Western world, INCPH might be responsible for 3%-5% of cases of portal hypertension.12 A histological review of 2500 autopsies demonstrated a prevalence of INCPH histological features of 3%. However, only 5% of these had evidence of portal hypertension.13 Concerning INCPH in the Western world, most studies were performed more than 15 years ago, enrolling patients for more than a decade earlier.14-16 Over the last decade, only two small cohort studies have been reported in the literature.6, 17 Based on these limited data, a male predominance with a median age of 40 years has been described. Exceptionally, INCPH has been reported in children.18
Many theories on the development of INCPH have been proposed, signifying limited understanding of the disease process. Theoretically, the etiology of INCPH can be divided in five categories: chronic infections, exposure to medication or toxins, genetic disorders, thrombophilia, and immunological disorders. Multifactorial etiology can also be encountered. One could speculate that the difference in worldwide prevalence of INCPH can be explained by a difference in genetic predisposition and area-specific diseases. In Western INCPH patients, a combination of disorders is often present.
INCPH has frequently been reported in association with immunological disorders.17, 19, 20 Various theories have been given to explain these associations. In patients with systemic sclerosis, a fibrogenetic process has been suggested as an etiological factor in the development of INCPH.21 Alternatively, in systemic lupus erythematosus patients, immunoglobulin (Ig) interference with prostacyclin formation has been designated to increase microthrombosis vulnerability.22 Immunoglobulin A (IgA) anticardiolipin antibody elevation, hypothetically leading to the obliteration of small vessels, has been demonstrated in the majority of celiac-disease–related cases of INCPH.19 Another immunological disorder with a high prevalence of INCPH is primary hypogammaglobulinemia. Malamut et al. demonstrated histological features of INCPH in 70% of these patients.23
Bacterial infection of the gut with repeated septic embolization and subsequent obstruction of small portal veins may be involved in the etiology of INCPH. This theory is supported by the high prevalence of INCPH in low socioeconomic areas with a high abdominal infection rate at birth and in early childhood.24 In addition, animal studies demonstrated that injection of Escherichia coli into the portal vein results in the development clinical and histological characteristics of INCPH.25
INCPH has been reported increasingly in patients with human immunodeficiency virus (HIV) infection.26-33 It remains a matter of debate whether a component of highly active antiretroviral therapy (HAART) or the presence of hypercoagulability plays a role in the development of HIV-related INCPH. Regarding the etiological role of HAART, prolonged exposure to didanosine has been assigned a potential role in its development. In a small cohort of HIV patients with cryptogenic liver disease, long-term didanosine treatment was observed in the majority of INCPH patients.27 Additionally, two recent case series reported long-term exposure of didanosine in 7 of 8 and 12 of 12 patients infected with HIV who had INCPH.32, 33 Its exposure has been shown to be associated with increased cardiovascular risk in HIV-infected patients, probably the result of a prothrombotic state secondary to the enhancement of proinflammatory mediators.34 Additionally, endothelial and mitochondrial damage of the portal system resulting from didanosine have been postulated in the pathophysiology of INCPH. Despite these hypotheses, it is difficult to conclude on the etiological role of didanosine, as the drug was widely used in the treatment of HIV in the past. Alternatively, a high prevalence of preexisting hypercoagulability (mainly protein S deficiency), possibly leading to vascular obstruction, has been reported in patients with HIV-related INCPH.26, 28, 29 This association remains controversial, as it has not been demonstrated consistently.32, 33
Medication and Toxins.
Several medications and chemicals have been alleged to cause INCPH. Among those, azathioprine, 6-thioguanine, and arsenic as Fowler's solution are the most frequently reported drugs associated with this disorder.35-37 Key et al. described the development of portal hypertension in five patients with chronic myeloid leukemia who were treated with busulphan and 6-thioguanine.38 However, because INCPH has also been associated with hematological diseases outside the setting of cytotoxic treatment, the association between this treatment and INCPH is not completely established.39 Currently, the most commonly used immunosuppressive drugs associated with the development of histological and clinical signs of INCPH are thiopurines (e.g., azathioprine and 6-mercaptopurine).40, 41 Although it is tempting to incriminate drug intake and chemical exposure as primary etiological factors, only a small minority of patients treated with the above-mentioned drugs or exposed to these chemicals develop clinical or histological signs of INCPH. It appears that an underlying susceptibility is needed to develop this disorder when exposed to the above-described agents.
Reports on the familial aggregation of INCPH and occurrence of its histological features in several congenital disorders (e.g., Adams-Oliver syndrome and Turner's disease) suggest a genetic background for this disorder.18, 42-45 The high prevalence of human leukocyte antigen (HLA)-DR3 positivity in these families supports an immunogenetic basis of this disorder.43
Hillaire et al. identified a 54% prevalence of prothrombotic disorders in a small patient cohort.6 An additional argument supporting the thrombophilia theory is the high prevalence and incidence of portal vein thrombosis in Western patients with INCPH. On the basis of clinical and histological data from INCPH patients, thrombophilia might be indicated as the underlying vulnerability necessary for the development of this disorder.46-49
Portal hemodynamics have been described to be different between INCPH and cirrhosis. A dual theory, implicating both increased splenic blood flow and intrahepatic obstruction, has been hypothesized regarding the development of INCPH (Fig. 1). Based on earlier studies, it has been speculated that the primary cause of INCPH is not related to hepatic abnormalities, but rather to an increased portal venous flow secondary to splenomegaly.50, 51 An overproduction of nitric oxide (NO), liberated in the sinus lining spleen cells, has been designated to be responsible for the dilatation of splenic sinuses and, subsequently, massive splenomegaly in INCPH patients.52 In these patients, liver specimens demonstrate normal histopathology. Observed disease remission after splenectomy supports the pathogenetic significance of splenomegaly in INCPH.53-55 In patients with more advanced disease, increased intrahepatic resistance resulting from obliteration of the portal venous microcirculation, presumably, would lead to a further elevation of portal hypertension. Thrombophilia, immunological disorders, and infections have been indicated as potential initiating lesions for portal venous obliteration.6, 49, 56, 57 However, because no supportive data are available, this theory remains an area of conjecture. An additional role has been attributed to endothelin-1 in the pathophysiology of INCPH. It has been speculated that an increased production of the latter increases vascular resistance and stimulates periportal collagen production.58
The majority of INCPH patients initially present with signs or complications of portal hypertension. In a large Indian study, 72% of patients with INCPH presented with gastrointestinal hemorrhage, whereas only a minority (14%) presented with splenomegaly.11, 59 In contrast, a low prevalence of upper gastrointestinal bleeding as an initial manifestation has been reported in Japanese and Western patients, of which the majority presented with splenomegaly or liver-test disturbances.6, 60 Compared to spleen enlargement in other causes of portal hypertension (e.g., liver cirrhosis and portal vein thrombosis), a massive, disproportionally large spleen is observed in patients with INCPH. In a large review on Indian INCPH patients, clinical splenomegaly was the most common initial symptom at the time of diagnosis (68.9%).10 In addition, 5.3% of these patients reported dragging pain caused by a huge mass. A minority of INCPH patients (30%) demonstrated impaired liver function at initial presentation in the context of gastrointestinal bleeding or in association with severe concurrent diseases. In general, liver function improved after controlling these associated conditions.6 Hepatic encephalopathy has rarely been reported in INCPH.61-63 Ascites has been described in 50% of INCPH patients.6 Comparable to liver failure, transient ascites occurs mainly in the presence of intercurrent conditions and mostly resolves after controlling the triggering events. Chronic ascites is described in association with renal failure and insulin-dependent diabetes mellitus in a minority of the patients. Until recently, hepatopulmonary syndrome was considered to be a rare complication in INCPH patients.64 However, two recent prospective studies reported a 10% prevalence of this pulmonary disorder in these patients.65, 66
Because there is no single test that can be regarded as the gold standard to diagnose INCPH, its diagnosis remains a challenge. Even in renowned hepatology centers, patients with INCPH are frequently misdiagnosed as having liver cirrhosis. Krasinskas et al. demonstrated that the majority of INCPH patients undergoing liver transplantation carried a pretransplantation diagnosis of cirrhosis.63
The initial assessment in patients with liver test disturbances or detected esophageal varices is typically performed with abdominal ultrasonography. Nodularity of the liver surface and thickening of the portal vein walls are sonographic features of INCPH (Fig. 2).10, 13, 46-48 However, these manifestations are not specific for INCPH and can also be observed in patients with liver cirrhosis. Recently, promising data have been published regarding discrimination between liver cirrhosis and INCPH with transient elastography.67 Mean liver stiffness in a large cohort of INCPH patients was 9.2 kPa, being significantly lower compared to the observed values in patients with liver cirrhosis (>14 kPa).68 As a result, the finding of liver stiffness values <14 kPa in the presence of clear signs of portal hypertension should raise the suspicion of INCPH.
Currently, liver biopsy remains essential in the diagnosis of INCPH. It is indispensible for the exclusion of liver cirrhosis, because, based on radiological examinations, INCPH patients are indistinguishable of cirrhotics. If liver cirrhosis and additional liver diseases known to cause portal hypertension histologically have been excluded, the pathologist has to look carefully for the discrete pathological characteristics of INCPH.
Macroscopic features in INCPH are mainly based on the examination of resection specimens from liver transplantation.46, 48, 49, 63, 69 The majority of these liver explants demonstrate organizing old thrombi (i.e., occluding or mural) in the large portal vein branches, nodular appearance, atrophy, and dysmorphy. In contrast, recent thrombi are rarely seen.70 In contrast, in some patients, gross appearance is normal.
Historically, INCPH has been classified in four different histological categories: idiopathic portal hypertension, NRH, partial nodular transformation (PNT), and incomplete septal cirrhosis.13, 24, 47, 48, 71-74 The presence of fibrotic portal tracts and thin fibrous septa in the absence of cirrhosis are pathological criteria for idiopathic portal hypertension.47, 73 In NRH, the parenchyma shows micronodular transformation, with central hyperplasia and an atrophic rim in the absence of fibrosis (Fig. 3).71 PNT is characterized by the presence of noncirrhotic, grossly visible parenchymal nodules located in the perihilar region of the liver around the large portal tracts.74 By definition, these nodules are larger than those in NRH, and diagnosis is only possible on resection specimens.75 Nodules macroscopically exhibiting PNT may show histological features similar to those in NRH (i.e., small hyperplastic nodules). Finally, incomplete septal cirrhosis is characterized by slender “incomplete” septal fibrosis that demarcates the parenchyma into conspicuous nodules with small hypoplastic portal tracts and hyperplastic hepatocytes.76, 77 Recently, this classification in different categories has been questioned.46 First, histopathological examination of whole livers from Western patients with INCPH demonstrated the concomitant presence of the different features in one specimen. Furthermore, pathological examination of livers resected at transplantation or autopsy failed to categorize the specimens according to the proposed classification because of the heterogeneity of the lesions.46, 48, 63, 78 As a result, in the Western world, INCPH is viewed as a single clinical entity with various pathological features, rather than separate clinicopathological entities. Although no pathognomonic histological findings exist in INCPH, frequently observed morphological features include the following: obliterative portal venopathy (luminal narrowing or obliteration of small portal venous branches accompanied by dense deposits of elastic fibers) (Fig. 4B); increased number of portal vascular channels; dilated portal veins herniating into the surrounding parenchyma (paraportal shunt vessels) (Fig. 4C); sinusoidal dilatation (megasinusoids); and periportal/perisinusoidal fibrosis.6, 13, 46, 47, 63, 76, 79, 80 Considering its high prevalence in INCPH liver specimens, obliterative portal venopathy is generally regarded as the primary lesion in the development of intrahepatic hemodynamical changes.6, 24, 81 According to Wanless, this obliteration of portal venules results in disturbed intrahepatic circulation and, subsequently, parenchymal remodeling, as observed in NRH and PNT (development of hepatocytic atrophy in the areas with reduced portal venous blood supply and compensatory hyperplasia in the best perfused areas).13 The additional morphological features of INCPH can be regarded as intrahepatic microcirculatory disturbances. For instance, the increased number of portal vascular channels and the paraportal shunt vessels (regarded the histological equivalent of the portal vein cavernoma) are believed to shunt blood from the obliterated portal segments toward unaffected tracts. Other morphological findings, however, are at variance with Wanless' obstructive portal vasculopathy theory. In the largest retrospective study on Western patients with INCPH to date, abnormal portal vessels were found in less than half of the cases. Furthermore, periportal and perisinusoidal fibrosis were more frequently observed in the absence, than in the presence, of portal vessel alterations. Therefore, Hillaire et al. suggested an alternative hypothesis, in which the causative process of the disease (a prothrombotic disorder) acts directly on the sinusoidal and portal vein wall, inducing fibrosis, obstruction, and secondary alterations in the architecture.6 Notwithstanding the occurrence or development of INCPH in patients with these histological features, a significant amount of patients with the described histological characteristics are observed in patients without clinical signs of portal hypertension.40, 82
Natural Course and Prognosis
Current data suggest that, despite liver function impairment occurring in the context of esophageal hemorrhage or infection, mortality of variceal hemorrhage in INCPH is significantly lower than that observed in cirrhotic patients.6, 16, 60, 76 None of the patients described by Hillaire et al. died from esophageal bleeding. As a result, isolated INCPH is regarded as a relative benign disorder (5-year survival of nearly 100%).24 Contrasting with this view, progression to liver failure (occurring late in disease course) requiring liver transplantation has been reported increasingly.17, 49, 63, 78 Cazals-Hatem et al. reported the development of severe liver failure in 7 of 59 patients with obliterative portal venopathy during a median follow-up of 8.6 years.49 Liver-function impairment and ascites in these patients can, possibly, be explained by a reduction in portal flow and, subsequently, atrophy of the peripheral hepatic parenchyma. In addition, the lack of compensatory arterial changes worsens ischemia and contributes to liver failure.83 The demonstration of obliterated large portal veins in explanted livers from INCPH patients transplanted because of liver failure supports this hypothesis.49 However, because no clear data are available, this hypothesis is speculative.
In comparison to patients with liver cirrhosis, a high incidence of portal vein thrombosis has been reported in patients with INCPH.6, 32, 84, 85 In patients with HIV-related INCPH, a substantially higher incidence of portal vein thrombosis (75%) has been documented,32, 85, 86 raising the possibility that HIV infection or its treatment may play a separate role in the development of portal vein thrombosis. A trend toward portal vein thrombosis being associated with poor prognosis has been reported.6 As a result, we believe that early diagnosis by regular screening of portal vein patency and, subsequently, the institution of anticoagulation therapy is strongly suggested. Considering the high incidence of portal vein thrombosis in INCPH, the occurrence of its histological features in patients with portal vein thrombosis, and the high prevalence of prothrombotic disorders in both conditions, it can be hypothesized that these two entities are different presentations of a single disorder.
The development of hepatocellular carcinoma in patients with INCPH remains a matter of debate. Notwithstanding, the reporting of liver cell atypia and pleomorphism in nodular regenerative hyperplasia liver specimens, a causal relationship between hepatocellular carcinoma and INCPH, has not been proven.39, 87 Nzeako et al. studied the association between NRH and hepatocellular carcinoma in 342 patients without cirrhosis.88 In the majority of the patients, the concurrence of NRH and hepatocellular carcinoma could be attributed to additional factors known to be associated with the development of NRH (e.g., portal vein thrombosis, chemotherapy, and radiotherapy), and the study failed to indicate this disorder as the underlying condition of hepatocellular carcinoma. As a result, hepatocellular carcinoma surveillance in patients with INCPH is not recommended.
Treatment and Prophylaxis of Variceal Gastrointestinal Bleeding.
Hemodynamics and, consequently, the management and prophylaxis of variceal bleeding in patients with INCPH are not entirely comparable to those in cirrhotic patients. Currently, scientific data on management and prophylaxis (i.e., primary and secondary) of variceal bleeding in INCPH patients are scarce. Nevertheless, we recommend to follow the guidelines of prophylaxis and management of cirrhotic variceal bleeding in patients with INCPH.89 Endoscopic sclerotherapy has been proven to be effective in controlling acute variceal bleeding of esophageal hemorrhage in 95% of INCPH patients.90 No scientific data have been published regarding endoscopic band ligation in these patients. However, considering the proven inferiority of sclerotherapy, compared with endoscopic variceal ligation in cirrhotic patients, the latter treatment is currently also regarded as the most appropriate endoscopic treatment in patients with INCPH. Despite the fact that data regarding combination of endoscopic treatment with vasoactive drugs and antibiotic prophylaxis are lacking, we recommend applying these treatments in INCPH patients, considering the effectiveness in cirrhotics. On the basis of Indian studies, emergency shunt surgery because of unmanageable bleeding was only required in 5% of cases with acute variceal bleeding.91 Despite the alleged safety and efficacy of shunt surgery, INCPH patients with uncontrollable hemorrhage are currently preferentially treated with transjugular intrahepatic portosystemic shunt (TIPS) because of its lower invasiveness.24 Taking into account the preserved liver function in these patients, the complications of this procedure observed in patients with cirrhosis (e.g., hepatic encephalopathy) are expected to be rare. However, no data are available. Concerning secondary prophylaxis of variceal bleeding in patients with INCPH, smaller studies have demonstrated a reduction of bleeding rate by endoscopic therapy.90, 92
Gastric varices are seen in nearly 25% of Indian INCPH patients.93 Portal hypertensive gastropathy is uncommon at initial presentation and is a rare cause of upper gastrointestinal bleeding.94 In patients with liver cirrhosis, nonselective β-blockers have been shown to reduce gastric mucosal blood flow and decrease recurrent bleeding in a randomized, controlled trial.95 In keeping with this, comparable treatment is applied in INCPH-induced portal hypertensive gastropathy.
Treatment of Splenomegaly and Hypersplenism.
Patients with INCPH have a massive splenomegaly, leading to increased portal venous flow and, subsequently, portal hypertension. As a result, splenectomy and partial splenic embolization have been demonstrated to decrease portal hypertension in these patients.96, 97 In selected patients with INCPH (e.g., abdominal discomfort or hypersplenism patients), these interventions can be regarded as effective therapeutic modalities.
Based on the high prevalence of thrombophilia and incidence of portal vein thrombosis in INCPH patients, several investigators have incriminated thrombosis of small intrahepatic portal veins as an important etiological factor in the development of this disorder.6, 30 Additionally, a trend toward poor prognosis has been reported in patients with INCPH who develop portal vein thrombosis.6 As a result, anticoagulation therapy has been proposed by several investigators to prevent disease progression and to maintain portal vein patency.6, 32, 98 However, considering the fact that gastrointestinal bleeding is the main complication of INCPH and the uncertain role of thrombophilia in the pathogenesis, this treatment is still a matter of debate and cannot be generally implicated until more solid data are present. Nonetheless, we believe that anticoagulation therapy must be considered in patients with underlying prothrombotic conditions and in patients who develop portal vein thrombosis.
Generally, patients with isolated INCPH have a normal liver function and the complications of portal hypertension can be managed successfully with endoscopic therapy and shunting. However, several reports describing liver transplantation in patients with INCPH have been published. The reported indications requiring liver transplantation in these patients were medical unmanageable portal hypertension, hepatopulmonary syndrome, hepatic encephalopathy, and progressive hepatic failure.49, 63, 78 Recently, Karsinskas et al. described a small cohort of INCPH patients treated with liver transplantation.63 The main indication for liver transplantation was medically unmanageable severe portal hypertension; a minority was listed because of hepatic encephalopathy. Notwithstanding the fact that resistant bleeding in INCPH patients should be treated with portosystemic shunting before considering the option of liver transplantation, only two patients underwent pretransplantation portosystemic shunting procedures (e.g., TIPS and mesocaval shunt). Presumably, the high frequency of cirrhosis misdiagnoses in these patients led to early referral for liver transplantation. To prevent unnecessary liver transplantation in these patients, early discrimination between cirrhosis and INCPH is extremely important. Based on small-sized cohorts (with limited follow-up), post-transplantation outcome in these patients is good and INCPH tends not to recur.63, 99, 100
Data on the etiology and management of INCPH are scarce, and currently applied diagnostic and therapeutic algorithms are based on personal experience or data from limited numbers of patients. The nomenclature concerning this clinical disorder has been ambiguent and has mainly depended on present histological features. To facilitate future studies and, subsequently, enhance our understanding of the disease, we propose INCPH as a uniform nomenclature for this disorder independent of the observed histopathological features. In Eastern patients, abdominally infectious disease has been incriminated as an important role in the development of INCPH; however, in Western patients, such a risk factor is lacking. Hypercoagulability may play an important role in INCPH. Despite the fact that data regarding treatment of variceal bleeding in INCPH patients are lacking, we recommend to follow the guidelines regarding cirrhotic variceal bleeding in these patients. In general, prognosis and survival of INCPH patients is good. However, liver failure might occur. Prospective multicenter cohort studies are needed to acquire reliable data regarding treatment and clinical outcome of this challenging disorder.
The authors are extremely grateful to Dr. P.E. Zondervan for critically reading parts of the manuscript for this article. The authors thank Dr. B. Liu for providing Fig. 1 and Dr. J. Verheij for providing Figs. 2 and 3.