Objective.We designed a longitudinal clinical database for autosomal recessive polycystic kidney disease (ARPKD), recruited patients from pediatric nephrology centers in the United States and Canada, and examined their clinical morbidities and survival characteristics. We initially targeted enrollment to children who were born and diagnosed after January 1, 1990, so as to capture a cohort that is representative of ARPKD patients born in the last decade. When a significant number of older ARPKD patients were also referred, we extended our database to include all patients who met our inclusion criteria, thereby allowing direct comparisons between a long-term survivor subset and a cohort that included both neonatal survivors and nonsurvivors.
Design.Patient entry into our database required either compatible histopathology or ultrasonographic evidence of enlarged, echogenic kidneys and the presence of at least 1 of the following additional criteria: a) biopsy-proven ARPKD in a sibling; b) biliary fibrosis based on either clinical or histopathologic evidence; c) no sonographic evidence of renal cysts in the parents (parents must be >30 years of age); or d) parental consanguinity, eg, first-cousin marriage. Clinical questionnaires (primary data form and follow-up data form) were developed to collect initial patient data and follow-up data at yearly intervals.
Results.Thirty-four centers provided clinical information for 254 patients and of these, 209 had sufficient data for analyses. When stratified by date of birth, 166 (79.4%) were born on or after January 1, 1990 (younger cohort) and 43 children (20.6%) were born before 1990 (older cohort). The gender distribution was equal in both cohorts. The median age at diagnosis was significantly later in the older cohort and no deaths were reported among these patients, suggesting that this group is biased toward long-term survivors. In the younger cohort, 74.7% of the patients are alive, with a median age of 5.4 years. In this group, 40.5% of patients required ventilation and 11.6% developed chronic lung disease. Hypertension was a common, but not universal finding in both cohorts. The relative risk for developing hypertension was higher in the older cohort, but the median age at diagnosis was significantly earlier in the younger cohort. Chronic renal insufficiency (CRI) was reported in ∼40% of patients with no significant difference in the relative risk between age groups. However, in the younger cohort, the median age at diagnosis was significantly earlier and the age of diagnosis of CRI and hypertension were significantly correlated. Clinically significant morbidities related to periportal fibrosis were more common in the older cohort. There was a trend toward increasing frequency of portal hypertension with age in both cohorts. Portal hypertension was not significantly correlated with either systemic hypertension or CRI.
Conclusions.The ARPKD Clinical Database represents the largest single cohort of ARPKD patients collected to date. Our initial data analysis provides several new clinical insights. First, in our subset of long-term survivors, ARPKD has a slower rate of disease progression, as assessed by age of ARPKD diagnosis, as well as age of diagnosis of clinical morbidities. Second, neonatal ventilation was strongly predictive of mortality as well as an earlier age of diagnosis in those who developed hypertension or chronic renal insufficiency. However, for infants who survive the perinatal period, the long-term prognosis for patient survival is much better than generally perceived. Third, although systemic hypertension and CRI were significantly correlated with respect to age of diagnosis, similar relationships with portal hypertension were not evident, suggesting that disease progression may have organ-specific patterns. Fourth, only a subset of patients may be at risk for developing clinically significant manifestations of periportal fibrosis. Based on these observations, the next challenges will be to determine how various factors, such as specific mutations in the ARPKD gene, PKHD1(polycystic kidney and hepatic disease 1), variations in modifying gene loci, modulation by as yet unspecified environmental factors, and/or gene-environment interactions contribute to the marked variability in survival and disease expression observed among ARPKD patients.
utosomal recessive polycystic kidney disease (ARPKD) is one of the most common pediatric renal cystic diseases, with an estimated incidence of 1 in 20 000 live births.1 The clinical spectrum is variable and depends on the age at presentation.2 The majority of cases are identified either in utero or at birth. The most severely affected fetuses have enlarged, echogenic kidneys and oligohydramnios attributable to poor fetal renal output. These fetuses develop the “Potter” phenotype, with pulmonary hypoplasia, a characteristic facies, and deformities of the spine and limbs. At birth, these neonates often have a critical degree of pulmonary hypoplasia that is incompatible with survival. Renal function, though frequently compromised, is rarely a cause of neonatal death. For those infants who survive the perinatal period, a wide range of associated morbidities can evolve, including hypertension, renal failure, and portal hypertension.2
Mutations at a single locus, PKHD1 (polycystic kidney and hepatic disease 1), are responsible for all typical forms of ARPKD.3 Based on the initial genetic studies, a reliable, haplotype-based diagnostic test has been developed for at-risk pregnancies.1 The recent identification of the PKHD1 gene by 2 groups, working independently and using different approaches,4,5 potentially provides the basis for gene-based diagnostic testing as well as for examining the contribution of specific PKHD1 mutations to the clinical variability observed among ARPKD patients. However, PKHD1 is a very large gene and the initial mutation detection rate is reported to be only ∼60%, suggesting that further analytic refinement must occur before gene-based analysis is robust for clinical use. In addition, Onuchic et al5 have demonstrated that the PKHD1 gene has a complex pattern of alternative splicing that is predicted to yield numerous putative protein products, some of which are membrane-bound and some are secreted. Therefore, the next challenge in deciphering the pathogenesis of ARPKD is to define the transcriptional and translational complexities of the PKHD1 gene to determine the role its protein product(s) play in the terminal differentiation of renal and biliary ductules.
Despite these recent molecular advances, the factors that modulate disease expression have yet to be defined and there remains a widespread pessimism about prognosis for ARPKD patients. To optimize the clinical decision-making regarding ARPKD patients, the natural history of ARPKD needs to be better characterized; the impact of recent advances in neonatal care and pediatric renal replacement therapy on ARPKD patient survival must be systematically evaluated, and clinical prognostic markers have to be developed. This study was undertaken to describe the baseline clinical features and survival characteristics of ARPKD patients drawn from tertiary care centers in the United States and Canada. These data provide the foundation for developing a longitudinal database for ARPKD patients.
PATIENTS AND METHODS
The ARPKD Clinical Database was initially designed to enroll patients diagnosed on or after January 1, 1990. The goal was to capture a cohort that was representative of ARPKD patients born and diagnosed in the last decade and to develop a longitudinal database to follow disease progression over time. When a significant number of older ARPKD survivors were also referred, we extended our database to include all patients who met our inclusion criteria.
Patients were eligible for enrollment if they had compatible histopathology based on renal biopsy or necropsy. Alternatively, patients with ultrasonographic evidence of diffusely enlarged, echogenic kidneys were eligible if they had the presence of at least 1 of the following additional criteria: a) patho-anatomic diagnosis in an affected sibling; b) absence of renal cysts in the ultrasound examination of both parents (parents must be >30 years of age); c) hepatic fibrosis based on either clinical or histopathological evidence; or d) parental consanguinity, eg, first-cousin marriage (associated with increased risk for an autosomal recessive disorder). Patients were excluded if they had autosomal dominant polycystic kidney disease (ADPKD), urinary tract malformations, or major congenital anomalies of other systems suggesting a diagnosis other than ARPKD.
Clinical questionnaires (primary data form and follow-up data form) were developed to collect initial patient data and follow-up data at yearly intervals. The primary data form recorded the dates of birth, diagnosis, last contact, and death (if applicable). Demographic characteristics such as gender and race were also recorded. A family history of polycystic kidney disease and parental consanguinity was requested. Perinatal data were collected, such as whether the diagnosis was made prenatally and if so, the diagnostic modality used, eg, sonography or genetic analysis. The primary data form recorded Apgar scores and neonatal interventions such as the use and modality of mechanical ventilation, eg, conventional versus high-frequency ventilation; administration of nitric oxide; treatment with extracorporeal membrane oxygenation or dialysis; and whether unilateral or bilateral nephrectomy had been performed. The presence of chronic lung disease (defined as the need for supplemental oxygen therapy) and date of diagnosis was recorded, as was whether the patient had evidence for growth retardation (defined as height <2 standard deviations for age). The date of diagnosis was requested for ARPKD-related renal morbidities, eg, hyponatremia (serum sodium <130 mg/dL), hypertension (defined as the date antihypertensive therapy was initiated), and chronic renal insufficiency ([CRI] defined as glomerular filtration rate [GFR] <75% normal adjusted for age, based on the estimated method of Schwartz6,7). The occurrence of urinary tract infections (UTIs; cystitis or pyelonephritis) was recorded and the abstractor was asked to note whether the diagnosis was culture-proven. The date of diagnosis was requested for ARPKD-related biliary morbidities, eg, portal hypertension (defined by sonographic evidence of hepatomegaly, splenomegaly, and directional reversal of portal vein flow), variceal bleeding, cholangitis, and hypersplenism. The date and results of sonographic and histopathologic studies of the kidney and liver were requested. Dates were recorded for initiation of therapeutic interventions such as antihypertensive medications (specific agent requested), recombinant growth hormone or erythropoietin, supplemental feeding through nasogastric or gastric tubes, dialysis (specific modality requested), variceal banding, portal vein shunting, or splenectomy. When applicable, the dates of kidney transplantation, the donor source (living-related donor vs cadaveric), and the status of the graft at last contact were requested. If appropriate, the date of liver transplantation was also recorded.
The follow-up data form recorded the date of last contact. When applicable, the date and cause of death was specified. The diagnosis dates were recorded for chronic lung disease and growth retardation. The date of diagnosis was requested for hyponatremia, hypertension, and CRI. The occurrence of UTIs (cystitis or pyelonephritis) was recorded and whether the diagnosis was culture-proven was noted. The date of diagnosis was requested for portal hypertension, variceal bleeding, cholangitis, and hypersplenism. The date and results of sonographic and histopathologic studies of the kidney and liver were requested. Dates were recorded for initiation of therapeutic interventions (eg, antihypertensive medications, recombinant growth hormone or erythropoietin, supplemental feeding, variceal banding, porto-systemic shunting, or splenectomy). When applicable, the dates of kidney transplantation, the donor source, and the status of the graft at last contact were requested. The date of liver transplantation was also recorded.
A pilot project was performed to assess the data forms for ease of data acquisition, potential ambiguities, and quality of data collected. A total of 12 to 15 patient records were collected from the Children’s Hospital of Alabama and 2 participating US pediatric nephrology programs (Fallon Clinic, Worcester, MA, and Connecticut Children’s Hospital, Hartford, CT). Each center completed the patient data forms and returned the forms and patient records to the University of Alabama at Birmingham. Parameters such as extraction time and data quality were compared between forms completed at the participating centers and those completed from the same medical records by S. A. Overstreet, RN (Clinical Research Coordinator, Children’s Hospital of Alabama). Primary data forms were revised and mailed to all members of the American Society of Pediatric Nephrology.
Those pediatric nephrologists who declined to participate were not contacted again. Follow-up telephone calls were made to pediatric nephrologists who did not respond to the initial mailing or who returned forms with incomplete data. Pediatric nephrologists who returned completed primary data forms were sent letters and follow-up data forms at yearly intervals for patient follow-up data. Participating pediatric nephrology centers received financial remuneration for contributing initial and follow-up patient data to the database.
To protect patient confidentiality, the contributing center was asked to provide patient initials and patient medical record number. A center-specific patient identifier was optional. Although the patient-specific information was required for initial database entry, this field was available only to the principal investigator and the biostatistician (R.A.D.) and was used to resolve data discrepancies with the research sites. No research personnel had access to any patient identifier such as patient name, initials, or medical record number that could allow an individual patient to be identified. Data security procedures and back-up were performed regularly by the data collection center.
All data forms were reviewed by the principal investigator (L.M.G-W.) for completeness and consistency of coding. The patient data were double-entered and the SAS COMPARE procedure in the SAS statistical package (SAS Institute, Inc, Cary, NC) was used to check the consistency of the 2 datasets.
Patient data were entered in a secure database and analyzed with the SAS statistical package version 8.2. Comparisons between categorical variables were analyzed with the χ2 test. Continuous variables were compared with the t test or, if the data were not distributed normally, the Wilcoxon rank sum test was used. Survival times were calculated from date of diagnosis to date of death or date of last contact for censored patients. Kaplan-Meier curves were constructed.9 Spearmann’s rank-order correlation coefficient r was used to examine the relationship between age at onset of various comorbitities by ventilation status.
Examination of multiple predictors of survival was conducted by Cox proportional hazards models. Because the proportional hazard assumptions did not hold across time, a state variable was created as a time-dependent covariate that takes into account patients who never developed the condition at the time of censoring. Prognostic factors included in the Cox model included CRI (1 = yes), hypertension (1 = yes), portal hypertension (1 = yes), gender (1 = female), and age at diagnosis of ARPKD (1: ≤7 days; 2: 8–365 days; 3: ≥365 days). For all analyses, a P value of <.05 was deemed statistically significant.
Patient Ascertainment and Cohort Characteristics
Thirty-four centers provided clinical information for 254 patients. Of these, 209 met our inclusion criteria and had sufficient data for analysis. When stratified by date of birth, 166 (79.4%) were born on or after January 1, 1990 (younger cohort) and 43 children (20.6%) were born before 1990 (older cohort).
As expected for an autosomal recessive disorder, there is an equal distribution of males and females in both age groups (Table 1). The entire cohort is dominated by whites, but other racial and ethnic groups are represented. Although prenatal diagnosis was rare in the older cohort (5.6%), it was reported for almost half (45.8%) of the younger cohort, with sonographic examination as the principal diagnostic modality used. The median age at diagnosis was significantly later in the children born before 1990 (72 days) than in the younger cohort (1 day; P = .0004). There were no deaths reported in the older cohort and the median age is 14 years at last follow-up (11.8–16.0 years; 25th–75th percentile), suggesting that this patient group is biased toward long-term survivors. In the younger cohort, 74.7% of the patients are alive, with a median age of 5.4 years (2.8–7.4 years; 25th–75th percentile). Among those children who died, respiratory failure and sepsis were the principal causes of death.
Neonatal morbidity was significantly less common in patients born before 1990, consistent with their older median age of diagnosis. Only 18.5% of the older cohort required mechanical ventilation and chronic lung disease was reported in 2.5%, whereas in the younger patients, 40.5% required ventilation and 11.6% developed chronic lung disease (Table2). Growth retardation was relatively common in both cohorts. Hyponatremia was reported in 26.5% of patients in the younger cohort, primarily in the neonatal period (median age: 16 days; Table 3). Culture-documented UTI was reported in 24.4% of the older cohort and 18.5% of the younger cohort, with females outnumbering males 3:1. In each cohort, there were no other significant differences in the frequency of ARPKD-associated morbidities when stratified by gender (data not shown).
Hypertension was a common, but not universal finding among the database patients. The relative risk (RR) for developing hypertension was higher in the older cohort (RR = 1.87; 95% confidence interval [CI]: 0.92, 3.82). However, the median age at diagnosis was significantly earlier in the younger cohort (16 days vs 240 days; P = .0003) and the age at diagnosis was quite variable (Table 3). Typically, multi-agent therapy was required for blood pressure control. The most commonly used agents were angiotensin-converting enzyme inhibitors, diuretics, and calcium channel blockers. Of note, hypertension was reported in 96% (44/46) of the patients who also reported hyponatremia.
CRI was reported in ∼40% of patients, with no significant difference in the RR between age groups. CRI occurred much earlier in the younger cohort (median age: 13.5 days vs 3.7 years in the older cohort; P = .0013) but the age at diagnosis was widely variable in both cohorts. Progression to end-stage renal disease was more common in the older cohort (RR = 1.77; 95% CI: 0.97, 3.23) for dialysis and 2.28 (95% CI: 1.26, 4.14) for renal transplantation. Among transplants, 16 were from living donors and 6 were cadaveric grafts.
Clinically significant morbidities related to periportal fibrosis were more common in the older cohort, (RR = 2.18; 95% CI: 1.27, 3.74) for portal hypertension and 2.12 (95% CI: 0.74, 4.81) for variceal bleeding. As with the renal-related morbidities, portal hypertension was diagnosed much earlier in the younger cohort (median age: 2.8 years vs 8.2 years in the older cohort; P = .0013) but the age at diagnosis was widely variable in both cohorts. The frequency of cholangitis was low and the likelihood was equivalent in the 2 cohorts. Though the sample size was small, we found evidence for a correlation between cholangitis and dilated intrahepatic bile ducts, as reported by others.8,9 Liver transplantation occurred more commonly in the older cohort (RR = 2.16) and all of these patients had coincident histories of variceal bleeding and/or cholangitis.
Correlation Among ARPKD Comorbidities
In the cohort of children born after January 1, 1990, neonatal ventilation was strongly associated with an earlier age of diagnosis in those who developed hypertension or CRI (Table 4). A similar correlation between mechanical ventilation and portal hypertension was statistically significant among females. Similar multivariate analyses were not possible in patients born before 1990, because only 5 of these patients required neonatal ventilation.
Among the patients born after 1990 who survived >30 days and developed CRI, 91% were also hypertensive (P < .0001). In this younger cohort, the age of diagnosis of CRI and hypertension were significantly correlated among both males and females (P < .0001, Spearman rank correlation coefficient; Table 5). In comparison, there was no significant relationship between development of systemic hypertension and portal hypertension (P = .92). The correlation between age of diagnosis of CRI and portal hypertension was not significant in the entire cohort (P = .08), but was borderline significant in males (P = .05).
Among 191 patients with recorded renal ultrasounds, 99 reported echogenic kidneys without cysts and 92 reported echogenic kidneys with small cortical cysts. Among patients with liver imaging studies (144 ultrasound, 6 computed tomography scan, and 2 other), 78 patients (52.7%) had normal liver images, 46 patients (31.0%) had echogenic livers, and studies in 24 patients (16.2%) revealed dilated intrahepatic bile ducts, consistent with Caroli’s disease. Among those with dilated ducts, 3 patients (13%) had had episodes of cholangitis.
Renal histopathology was available for 46 patients initially identified. Forty-two patients had dilated collecting ducts, 2 had glomerular dilatation, and 2 had dysplastic elements. More detailed analysis of these latter 4 patients, including review of their pathology reports, raised doubts about the diagnosis of ARPKD and these patients were removed from the database for further follow-up. Liver histopathological analyses were reported for 45 patients. In 43 there was evidence of biliary dysgenesis and two were normal. Given that ARPKD is invariably associated with biliary dysgenesis, these latter 2 patients were removed from the database.
No deaths were reported in the cohort of patients born before 1990. In the subset of children born after January 1, 1990, overall patient survival was 85.8% at 1 month, 78.6% at 1 year, and 74.6% at 5 years (Table 6). The highest mortality rate occurred within the first month of life (Fig 1) with 21 (58%) of the 36 deaths occurring in this interval. These deaths apparently were related to respiratory insufficiency, as mechanical ventilation in the neonatal period was a strong negative predictor of long-term survival. Among all deceased patients, 34/36 (94.4%) required ventilation as neonates (P < .0001). Of those patients who survived the first month of life, 91.7% were alive at 1 year and 87% were alive at 5 years.
In addition to the major effect of neonatal respiratory insufficiency on survival outcomes, both age of diagnosis and occurrence of CRI were significant predictors of mortality (Table 7). For example, patients with CRI were 2.4 times more likely to die compared with patients with normal renal function, controlling for the other factors listed. There was no significant effect of either gender or hypertension on mortality. Although portal hypertension appears to be an important predictor (hazard ratio [HR] = 5.87) of mortality, the significance of this effect was marginal (P < .1), perhaps resulting from the small number (n = 24) of affected patients.
Although several previous series have examined ARPKD-related morbidities and survival characteristics,8–13 most of these studies were conducted before the 1990s, before the advent of newer therapeutic modalities for the neonatal and pediatric management.
We therefore sought to reassess ARPKD mortality rates and the frequencies of ARPKD-related morbidities in a more recent cohort drawn from US and Canadian pediatric nephrology centers. We initially targeted children who were born and diagnosed after January 1, 1990, for recruitment to the ARPKD Clinical Database. We acknowledge that this design is inherently biased toward patients who are referred to pediatric nephrologists and thus, are more likely to have survived the first 24 hours of life. Therefore, this database is not optimal for accurately assessing the neonatal incidence of ARPKD, nor the associated mortality rate. Such analyses would require comprehensive reporting from perinatology groups that cover >20 000 live births per year.1 However, we expected that our database design would capture a cohort of both neonatal survivors and nonsurvivors and thus be somewhat representative of ARPKD patients born in the last decade. When a significant number of older ARPKD patients were also referred, we extended our database to include all patients who met our inclusion criteria. This allowed us to compare the clinical characteristics of a long-term survivor subset with that of a cohort that included both neonatal survivors and nonsurvivors.
In both cohorts, ARPKD is expressed primarily in whites. However, this disorder is also observed in children of other racial and ethnic origins, a relatively underappreciated point in the literature. Aside from the strong female predominance among patients reporting UTIs, we did not detect gender differences in disease diagnosis, frequency of ARPKD-associated morbidities, or mortality. Therefore, our data do not support the findings previously reported by Zerres et al,9 where girls had a statistically significant higher mortality rate and more severe disease expression with higher frequency of UTIs, growth retardation, impaired renal function, and progression to end-stage renal disease.
Our analyses indicate that the cohort of patients who were diagnosed before 1990 (older cohort), was strongly biased toward long-term survivors. No deaths were reported among these patients. They had a later age of ARPKD diagnosis as well as diagnosis of both renal and biliary-related morbidities. These data demonstrate a previously reported,12,14,15 but relatively underrecognized fact, ie, long-term survival does occur in ARPKD and this subset of patients has a slower rate of disease progression. This patient subset provides a unique resource for investigating how various biological, genetic, and/or environmental factors modulate ARPKD-related disease expression. The resulting insights should help define the molecular mechanisms involved in ARPKD pathogenesis and establish a framework for designing therapeutic interventions that attenuate expression of ARPKD-associated morbidities.
Among patients diagnosed after January 1990 (younger cohort), the 1-year and 5-year survival rates are consistent with previous reports (Table 8). Although these data suggest that recent therapeutic advances in neonatal and pediatric management have not significantly impacted the clinical outcome in ARPKD patients, it is important to emphasize that the patient outcomes in most of these studies is better than generally perceived by obstetricians, perinatalogists, and pediatricians. In our database, which represents the largest single ARPKD cohort reported to date, the 1- and 5-year survival was ∼90% for those neonates who survived the first month of life.
Approximately 40% of ARPKD neonates required mechanical ventilation. These infants had a higher mortality rate than nonventilated infants and in those patients who developed hypertension or CRI, the age of diagnosis was significantly earlier. This ventilation requirement likely reflects more significant pulmonary hypoplasia, which in turn may suggest more severe renal cystic disease. We therefore propose that clinicians regard neonatal ventilation as a prognostic marker that portends lower patient survival, as well as the more rapid development of renal-associated morbidities.
Growth retardation was relatively common in both of our cohorts. Konrad et al16 reported similar data in ARPKD patients and ascribed the growth retardation, particularly in girls, to the rapid deterioration in their renal function. However, in our cohorts, delayed growth could not be readily attributed to any specific causal factor(s), as only some of these patients had coincident chronic lung disease and/or CRI.
Hyponatremia was reported in 26% of the younger cohort, primarily among neonates. The vast majority of these patients were born after 30 weeks’ gestational age and thus, the prevalence of hyponatremia in our cohort was significantly higher than that previously reported for age-matched infants. For example, Al-Dahhan et al17 observed hyponatremia in 13% of those born at 30 to 32 weeks and none of those born after 32 weeks. We note that in our cohort there was significant overlap between the median age of diagnosis of hyponatremia and CRI, making it somewhat difficult to define the etiologic factors contributing to the hyponatremia. However, hyponatremia was not reported among children older than 2 months of age who had coincident CRI.
In our series, as well as previous reports, systemic hypertension was the most common comorbidity associated with ARPKD. The pathogenesis of ARPKD-related hypertension remains unclear. Although there is compelling evidence that a renin-mediated mechanism underlies ADPKD-related hypertension,18 Kaplan et al19 have reported that hypertensive ARPKD patients have low plasma renin levels and expanded intravascular volume, particularly those with concomitant hyponatremia.19 This association suggests that ARPKD-related hypertension may develop as the result of a defect that dysregulates sodium reabsorption in the ectatic collecting ducts. The current study provides circumstantial evidence in support of this speculation, as 96% of the patients reporting early-onset hypertension also had hyponatremia. Unfortunately, plasma renin values were not available for these infants. It must be noted, however, that other ARPKD reports8,9 have not corroborated the association between hyponatremia, intravascular volume expansion, and low-renin hypertension and the current standard-of-care for management of ARPKD-related hypertension continues to rely on angiotensin-converting enzyme inhibitors and calcium channel blockers. Further investigation will be required to dissect the molecular basis for hypertension in these patients and to guide the development of more targeted therapy.
In patients born after 1990, we found a significant correlation between hypertension and CRI, with respect to the age at diagnosis. Similar associations between systemic hypertension and poorer actuarial renal survival have been reported for numerous other renal disorders, including ADPKD.20 However, it has been difficult to establish whether hypertension is a marker of disease severity or a progression factor that accelerates a decline in renal function. Unfortunately, our database design does not permit direct analysis of the potential cause and effect relationship between hypertension and CRI in our ARPKD cohort.
In previous studies, the assessment of periportal fibrosis varied widely in terms of prevalence rates and specific parameters reported (Table 8). In the longest longitudinal study, Roy et al12 observed hepatomegaly in 83% and splenomegaly in 68%, whereas only 23% of patients had bleeding from esophageal varices and 20% were found to have hypersplenism. In the cohort reported by Zerres et al,9 46% of patients presented with clinical signs of periportal fibrosis defined as sonographic evidence of hepatomegaly or splenomegaly, or esophageal varices detected by radiographic or endoscopic examination. Gagnadoux et al8 observed clinically significant portal hypertension requiring porto-systemic shunting in 4 (13%) of the 30 patients who survived the first year of life. Hepatomegaly and/or sonographic evidence of intrahepatic biliary duct dilatation was found in ∼50% of these patients. In a recent study by Capisonda et al,13 sonographic evidence of portal hypertension, characterized by visualization of the portal vein, splenomegaly, directional reversal of portal vein flow, and the presence of varices, was noted in 10 (37%) of the 27 surviving patients. Bleeding esophageal varices occurred in 3 patients and all of them required sclerotherapy to control hematemesis. Hepatomegaly was documented in 16 (52%) patients and splenomegaly in 11 (36%).
Portal hypertension in the current study was defined by sonographic criteria, eg, hepatomegaly, splenomegaly, and directional reversal of portal vein flow, and reported in 35% of the older cohort (median age: 8.2 years) compared with 15% of the younger cohort (median age: 2.8 years). Although there is considerable variability in the assessment parameters used among various studies reported to date, our data are consistent with previous studies in suggesting a trend toward increasing frequency of portal hypertension with age. However, we observe that the median age of diagnosis in each cohort is significantly younger than the median patient age per cohort. This new, previously unappreciated finding suggests that only a subset of patients may be at risk for developing clinically significant manifestations of periportal fibrosis. If this observation is confirmed in continued longitudinal surveillance of our cohorts, the next challenge will be to determine how various genetic and/or environmental factors contribute to the progression of biliary fibrosis in ARPKD patients. These investigations will have both prognostic and therapeutic implications.
Only a few previous studies have examined the correlation between renal-related and biliary-related morbidities. Zerres et al21 proposed that the extent of periportal fibrosis increased directly with prolonged survival in patients with milder renal disease. In contrast, Gagnadoux et al8 did not find any relationship between renal and biliary involvement. Data from our cohort is more consistent with Gagnadoux et al, as we found no correlation between the development of systemic hypertension and portal hypertension. The age of diagnosis of CRI and portal hypertension also was not significantly correlated. Although a relationship between these comorbidities was suggested in males, the sample size was small. Our longitudinal study will continue to evaluate whether the progression of renal disease and periportal fibrosis are related in some patient subsets or whether they proceed as independent processes. These data will have important implications for the development of clinical prognostic markers and the design of therapeutic interventions that target organ-specific disease progression.
Our ARPKD Clinical Database has enrolled >200 patients and thus, represents the largest single cohort of ARPKD patients collected to date. Continued patient accrual and longitudinal follow-up should increasingly facilitate the development of prognostic markers for patient survival and disease progression. The next challenges will be to determine how various factors, such as specific mutations in the PKHD1 gene, variations in modifying gene loci, modulation by as yet unspecified environmental factors, or gene-environment interactions contribute to the marked variability in survival and disease expression observed among ARPKD patients. The well-characterized patient cohort described in our database offers a critical clinical resource for these investigations.
This work was supported by a Polycystic Kidney Disease Foundation special grant-in-aid to Dr Lisa M. Guay-Woodford.
We thank Sandra Overstreet, RN, and Carthenia Jefferson, RN, for assistance with this study.
The authors would also like to acknowledge the members of the ARPKD Clinical Database Steering Committee: Ellis D. Avner (Case Western Reserve University, Cleveland, OH), Barbara R. Cole (Washington University, St Louis, MO), Bernard S. Kaplan (University of Pennsylvania, Philadelphia, PA), Clifford Kashtan (University of Minnesota, Minneapolis, MN), Majid Rasoulpour (University of Connecticut, Hartford, CT), and Aileen B. Sedman (University of Michigan, Ann Arbor, MI).
The following institutions generously contributed patient data to the North American ARPKD Clinical Database: Children’s Hospital of Alabama (Birmingham, AL); University of Alberta (Edmonton, Alberta, Canada); Arkansas Children’s Hospital (Little Rock, AR); St Joseph’s Hospital (Phoenix, AZ); Stanford University School of Medicine (Stanford, CA); Connecticut Children’s Medical Center (Hartford, CT); Yale University School of Medicine (New Haven, CT); Children and Infants’ Diagnosis Center (Margate, FL); Nemours Children’s Clinic (Orlando, FL); Medical College of Georgia (Augusta, GA); Children’s Kidney Center of Illinois (Chicago, IL); Children’s Hospital of Boston (Boston, MA); Fallon Clinic (Auburn, MA); Johns Hopkins University (Baltimore, MD); C. S. Mott Children’s Hospital-University of Michigan (Ann Arbor, MI); University of Minnesota (Minneapolis, MN); St Louis Children’s Hospital-Washington University (St Louis, MO); Duke University Medical Center (Durham, NC); Children’s Mercy Hospital (Kansas City, MO); Babies and Children’s Hospital/Columbia-Presbyterian (New York, NY); Children’s Hospital Buffalo (Buffalo, NY); Mount Sinai Medical Center (New York, NY); New York Medical College (Valhalla, NY); North Shore University Hospital (Manhasset, NY); Rainbow Babies & Children’s Hospital (Cleveland, OH); Children’s Medical Center (Dayton, OH); Ohio State University (Columbus, OH); The Hospital for Sick Children (Toronto, Ontario, Canada); University of Ottawa (Ottawa, Ontario, Canada); Children’s Hospital of Philadelphia (Philadelphia, PA); M. S. Hershey Medical Center-Penn State College of Medicine (Hershey, PA); Le Bonheur Children’s Hospital (Memphis, TN); Children’s Hospital of the King’s Daughters (Norfolk, VA); and Robert C. Byrd Health Sciences Center (Charleston, WV).
- Received September 16, 2002.
- Accepted November 12, 2002.
- Reprint requests to (L.M.G-W.) Division of Genetic and Translational Medicine, University of Alabama at Birmingham, Kaul 740, 1530 Third Ave S 19th St, Birmingham, AL 35294. E-mail:
- ↵Guay-Woodford L. Autosomal recessive disease: clinical and genetic profiles. In: Watson M, Torres V, eds. Polycystic Kidney Disease. Oxford, United Kingdom: Oxford University Press; 1996:237–267
- ↵Schwartz G, Haycock G, Edelmann C, Spitzer A. A simple estimate of glomerular filtration rate in children derived from body length and plasma creatinine. Pediatrics.1976;58 :259– 263
- ↵Gagnadoux M-F, Habib R, Levy M, Brunelle F, Broyer M. Cystic renal diseases in children. Adv Nephrol.1989;18 :33– 58
- ↵Al-Dahhan J, Haycock G, Chantler C, Stimmler L. Sodium homeostasis in term and preterm neonates: I. Renal aspects. Arch Dis Child1983;58 :335– 342
- ↵Chapman A, Gabow P. Hypertension in autosomal dominant polycystic kidney disease. Kidney Int (Suppl).1997;61 :S71– S73
- ↵Zerres K, Volpel MC, Weiss H. Cystic kidneys: genetics, pathologic anatomy, clinical picture, and prenatal diagnosis. Hum Genet1984;68 :l04– l35
- Copyright © 2003 by the American Academy of Pediatrics