The Management of Pediatric Renovascular Hypertension: A Single Center Experience and Review of the Literature
Introduction: Renal artery occlusive disease is poorly characterized in children; treatments include medications, endovascular techniques, and surgery. We aimed to describe the course of renovascular hypertension (RVH), its treatments and outcomes.
Methods: We performed literature review and retrospective review (1993-2014) of children with renovascular hypertension at our institution. Response to treatment was defined by National High Blood Pressure Education Program Working Group on High Blood Pressure in Children and Adolescents at most-recent follow-up.
Results: We identified 39 patients with RVH. 54% (n=21) were male, with mean age of 6.93 ± 5.27 years. Most underwent endovascular treatment (n=17), with medication alone (n=12) and surgery (n=10) less commonly utilized. Endovascular treatment resulted in 18% cure, 65% improvement and 18% failure; surgery resulted in 30% cure, 50% improvement and 20% failure. Medication alone resulted in 0% cure, 75% improvement and 25% failure. 24% with endovascular treatment required secondary endovascular intervention; 18% required secondary surgery. 20% of patients who underwent initial surgery required reoperation for re-stenosis. Mean follow-up was 52.2 ± 58.4 months.
Conclusions: RVH treatment in children includes medications, surgical or endovascular approaches, with all resulting in combined 79% improvement in or cure rates. A multidisciplinary approach and individualized patient management are critical to optimize outcomes.
Key words: renovascular hypertension, renal artery stenosis, hypertension, pediatrics
Pediatric hypertension has been previously determined to be one of the strongest predictors of adult hypertension, which, if uncorrected, leads to sequelae including myocardial infarction, stroke and encephalopathy . Renovascular hypertension (RVH) is defined as high blood pressure resulting from a lesion reducing blood flow to part or all of one or both kidneys with subsequent alteration of the renin-angiotensin mechanism . This rare, but serious, cause of high blood pressure in children accounts for approximately 10% of all causes of secondary hypertension . The most common etiology of renal artery stenosis (RAS) in children is fibromuscular dysplasia (FMD), and other causes include syndromes (neurofibromatosis type 1 (NF1) , Williams syndrome , tuberous sclerosis ) and vasculitides (Takayasu arteritis , polyarteritis nodosa , Kawasaki disease ).
Unlike hypertension in adults, in which atherosclerotic lesions result in non-correctable disease in 70-80% of patients , RVH in children has the potential to be cured by medical, endovascular, or surgical interventions . Medical therapy alone is typically undertaken in children for whom surgical or radiological interventions are not determined to provide further benefit or are not feasible . This non-invasive treatment however, may present challenges related to patient adherence and adverse drug effects. More invasive interventions include percutaneous transluminal angiography (PTA) with balloon dilation or open surgical techniques in which the affected stenotic portion of the renal artery is either removed or bypassed or the entire affected kidney is removed. These interventions, while often providing initial effective control of patient blood pressure, may not always be durable . The goal of all management approaches in children includes lifelong preservation of renal function with restoration of renal perfusion .
This study aimed to determine the presentation, course, treatment and subsequent outcomes of various treatment modalities for RVH in the pediatric population at a single institution. An extensive literature review was also performed evaluating the presentation, etiologies, diagnostic modalities and treatments of children with RVH.
After Institutional Review Board (IRB) approval, we retrospectively reviewed the medical records of consecutive patients diagnosed with RVH from 1993 – 2014 at Cincinnati Children’s Hospital Medical Center. The radiology database of all patients at our institution was queried for those who demonstrated possible partial or complete renal artery occlusion on imaging, and 717 imaging studies were identified. The electronic medical records of all patients were reviewed to identify those with RVH on digital subtraction angiography or gross examination during surgery and duplicates were removed. Of the 52 patients with verified RVH on imaging or at surgery, 39 had full medical records regarding diagnosis, treatment and outcomes available for evaluation and were included in the study. Variables assessed included patient age at diagnosis, gender, past medical history, presenting symptoms, diagnostic laboratories and imaging including ultrasound, echocardiography, MRI and CT scanning, treatment including endovascular and operative details, complications, and outcomes.
The National High Blood Pressure Education Program Working Group on High Blood Pressure in Children and Adolescents  was utilized to assess response to interventions. Blood pressure response accounted for age and gender and was defined as:
(1) cured (blood pressure reduced to below the 95th percentile without requiring antihypertensive medications);
(2) improved (blood pressure reduced below 95th percentile while still requiring antihypertensive medications or diastolic blood pressure (DBP) reduced by more than 15% of pre-intervention level); or,
(3) failed (hypertension persisted despite antihypertensive medications with less than a 15% decrease in DBP from pre-intervention level). Blood pressure at most recent follow-up was obtained from clinical outpatient records.
A literature review of renovascular hypertension in children was performed utilizing PubMed, including English studies of children < 18 years of age. Studies that were not in English were excluded.
A cohort of 39 patients with RVH was assessed and was comprised of 21 males and 18 females. The mean age at presentation was 6.93 ± 5.27 years (range, birth – 17 years), at diagnosis of RVH was 7.91 ± 5.22 years (range, birth – 17 years), and at initial treatment was 8.11 ± 5.31 years (range, 5 weeks – 17 years). Thirty-eight patients were diagnosed at our institution, and one patient was diagnosed at a referring hospital prior to presentation to Cincinnati Children’s Hospital Medical Center. Hypertension (>95th percentile for age and gender) was present in all patients at time of presentation. Most patients (n=30, 77%) presented with an incidental elevated blood pressure measurement at outpatient clinic or hospital visit for other complaint. Other presenting signs and symptoms included hypertensive encephalopathy, proteinuria/hematuria, fatigue, abdominal bruit and emesis, as shown in Table 1. Of patients with hypertensive encephalopathy, most complained of headache (n=6, 55%), visual changes (n=2, 18%) or seizures (n=2, 18%). Other neurologic symptoms included mental status changes (n=1, 9%), tinnitus (n=1, 9%), intracranial hemorrhage (n=1, 9%), and facial nerve paralysis (Bell’s palsy) (n=3, 27%). Heart failure was diagnosed in 3 children.
In patients with reported etiologies (n=20), the majority of patients had RVH due to FMD (n=7, 35%) or syndromic causes (n=8, 40%), as shown in Table 2. Syndromes presenting in patients included NF1 (n=3), Williams syndrome (n=2), Sturge-Weber syndrome (n=1), von Willebrand’s Disease (n=1) and Bartter syndrome (n=1). Diffuse vascular disorders were diagnosed in 3 patients, including Moyamoya disease and Takayasu arteritis in one, steno-occlusive splanchnic arteriopathy in a second and mid-aortic syndrome in a third. Single-vessel extra-renal disorders were seen in five patients, with coarctation of the aorta and aortic stenosis in two patients each and carotid occlusion in one patient.
Several diagnostic modalities were utilized in our cohort. Doppler ultrasound was the most common non-invasive diagnostic imaging technique (n=26, 67%) with magnetic resonance angiography (MRA; n=12, 31%) and computed tomographic angiography (CTA; n=11, 28%) being less commonly utilized. All patients who underwent MRA or CTA imaging examinations had undergone a prior Doppler ultrasound study. Ultrasonography was noted to be the least sensitive modality, with 54% (n=14) of patients with RVH displaying normal kidneys, and only 46% (n=12) showing evidence of renal artery stenosis, including increased renal artery resistive index, pulsus parvus et tardus waveform, or increased peak systolic velocity. Only one patient who underwent CTA and three patients who underwent MRA were reported to have normal kidneys on imaging. Most patients (n=33, 85%) underwent catheter-based diagnostic angiography for definitive diagnosis. Figure 1 displays a typical catheter-based diagnostic angiogram of a patient with renal artery stenosis.
Renal artery stenosis was unilateral in most patients (n=29, 74%). One patient had stenosis in a solitary kidney. Of the 9 patients who had bilateral lesions, three had diffuse narrowing of other vessels on angiography, including the aorta, common hepatic artery and superior mesenteric artery (SMA)—two of these patients had renal artery stenosis as a part of a syndrome or vasculitis (Moyamoya / Takayasu arteritis and steno-occlusive splanchnic arteriopathy). Overall in our cohort, the stenosis typically involved one discrete short lesion (n=24, 62%) and involved the main renal artery (n=26, 67%).
Treatment modalities for pediatric patients with RVH varied, with endovascular management by balloon angioplasty (Figure 1) being the most commonly utilized (n=17, 44%). Stenting was not utilized, and no immediate complications were observed following endovascular intervention. Primary surgical intervention (n=10, 26%) and medication-only treatment (n=12, 31%) were less common approaches to management (Figure 2). Primary surgical management included unilateral (n=4) and bilateral (n=2) renal artery bypass, unilateral (n=1) and bilateral (n=1) renal artery reimplantation, and nephrectomy (n=2). Three patients (18%) who underwent initial endovascular management required repeat angioplasty. One of these patients had Bartter syndrome and developed a contralateral renal artery stenosis. The other two were asyndromic and developed marked focal narrowing of their main renal arteries. Three additional patients (18%) underwent surgical management after initial endovascular treatment: unilateral renal artery bypass (n=1), nephrectomy (n=1), and interposition graft with eventual reoperation for nephrectomy (n=1). Likewise, two patients (20%) who underwent initial surgical treatment required repeat operative treatment for recurrent or new stenosis in the setting of refractory hypertension; both patients underwent nephrectomy as definitive surgical management due to non-reconstructable renal artery disease. Table 3 displays blood pressure values at presentation and at 1 year post-treatment, as well as outcomes of patients at most recent follow-up utilizing the National High Blood Pressure Education Program Working Group on High Blood Pressure in Children and Adolescents guidelines. Primary surgical management resulted in the highest rate of cure (30%), with endovascular treatment curing 18% of patients, and medication-only treatment resulting in no patients with cured RVH. Mean follow-up was noted to be 52.2 ± 58.4 months. When combining all treatment modalities, 79% of patients had improved or cured hypertension at most recent follow-up. No patient has developed renal failure during the follow-up period.
Sustained high blood pressure occurs in 1-2% of children, with secondary causes occurring more often than primary idiopathic hypertension [16-18]. Of patients with secondary hypertension, RVH is a significant etiology, affecting approximately 10% of patients , after renal scarring and glomerular disease . RVH differs from other causes of hypertension in that it has potential for cure with medical, endovascular or surgical interventions. In this study, we assessed the clinical presentation, diagnosis, treatment and outcomes of 39 patients with heterogeneous renal artery lesions resulting in pediatric RVH, and provide one of the largest series of pediatric RAS to date, behind Stanley et al [19, 20] Lacombe et al  , Sandmann el al  and Kari et al . We also review the current literature regarding renovascular hypertension in children.
RVH should be suspected in children with high blood pressure or secondary symptoms of high blood pressure. In the current study, incidental hypertension, detected at an outpatient clinic or hospital visit for an asymptomatic child was the most common presenting symptom, occurring in 77% of children. Other studies have also reported 26-70% of RVH to be incidentally detected in asymptomatic children. In these patients, elevated blood pressure is often not amenable to treatment with single anti-hypertensive agents and may progress to more serious sequelae, such as cerebrovascular incidents and hypertensive encephalopathy . Some have reported that older children and adolescents are most likely to present without symptoms [19, 23] whereas younger children are most likely to have more severe neurological (seizures) and cardiac (left ventricular hypertrophy and congestive heart failure) symptoms or lethargy and poor growth. Overall, behind an asymptomatic presentation, symptoms in published series include headache, early fatigue, congestive heart failure and intermittent claudication. More severe presentations such as cerebrovascular accidents and severe deterioration of renal function are common in young patients. Left ventricular hypertrophy was detected in 18- 66% of patients in the largest published series [22, 23]. In our series only 3 (8%) of patients had cardiac dysfunction and 33 % demonstrated neurologic manifestation. The delay, and often difficulty, of diagnosis of hypertension in the child may partially account for this. Too frequently, elevated blood pressure in the young child is attributed to inaccurate measurement.
Whereas the main etiology of renovascular hypertension in the adult population is atherosclerotic disease, the etiology of RVH in the pediatric population varies. The two main causes reported in the literature are fibromuscular dysplasia (FMD) and Takayasu arteritis. Tullus et al noted that the majority of children in North America and Europe are diagnosed with FMD. However, Takayasu arteritis is the predominate diagnosis in Asia and South Africa . This has been corroborated by other published RVH series, as studies from China and Turkey reported Takayasu arteritis in 72%  and 60%  of RVH patients, respectively. Other syndromic associations include Neurofibromatosis 1 (NF1) [19, 23], Williams syndrome [23, 26], and other arteritides such as polyarteritis nodosa and Kawasaki disease .
Similar to other North American series [11, 14, 20, 23, 24], FMD was the most common cause of RVH in our series when excluding the combined syndromic causes. An additional 21 patient had no definitive reported etiology for their hypertension although we suspect that the majority of these were due to FMD. This likely stems from the difficulty in diagnosing FMD. FMD is a non-atherosclerotic, non-inflammatory angiopathy of unknown cause affecting medium-sized arteries.FMD involves dysplasia or hyperplasia of a portion of the arterial wall, with medial and perimedial FMD accounting for the majority of disease [27, 28]. Classically, on angiography, it has been described with the “string of beads” appearance [29, 30], characteristic of multifocal FMD. Since patients with the focal type of FMD were younger at diagnosis, the actual rate of observance of the “string of beads” manifestation is unknown and frequently multiple modalities are required for diagnosis.
Other common causes in our cohort included syndromes (NF-1, Williams syndrome, Moyamoya) and vasculitides (Takayasu arteritis). Clinicians treating patients with these syndromes and vascular disorders should have a high index of suspicion for RVH, as previous studies have reported that up to 58% of patients with NF1 have RVH [17, 31], and a study from South Africa has reported 89% of children with RVH to have the primary diagnosis of Takayasu arteritis .
Mid-aortic syndrome (MAS) is another etiology of RVH and refers to the localized narrowing of the distal thoracic/abdominal aorta and usually involves the renal arteries and visceral branches, regardless of etiology [33, 34]. The etiology can be congenital or acquired but frequently is due to inflammatory conditions of the aorta such as Takayasu and Neurofibromatosis . As many as 91% of patients with MAS have RVH  and anywhere from 20-48% of RVH can be attributed to MAS [5, 36]. The presentation of MAS is often related to the RVH but also frequently related to stenosis of other visceral and extremity arteries such as bilateral lower extremity claudication and abdominal migraine.
Tullis et al recently described a diagnostic algorithm for pediatric patients with RVH . In this approach, physical examination including repeating blood pressure measurement is followed by non-invasive imaging (such as Doppler ultrasonography), a trial of anti-hypertensive medication, and finally, angiography, if response to medication-only treatment is poor . In evaluation of pediatric patients with hypertension, although RVH was suspected in nearly all, some points in the history and exam were particularly concerning. Secondary findings such as cardiac failure and cerebral symptoms were common in the younger populations in many of the series [19, 22-24]. Other findings that should generate concern for RVH include association with a syndrome that is high-risk for vascular disease such as signs of vasculitis, NF1, Williams syndrome. These should trigger an early diagnostic evaluation for RVH.
When the suspicion is high for RVH, digital subtraction angiography is the gold standard . Although the role of non-invasive imaging in children with RVH is not well established, these studies have been described as sole diagnostic modalities in the diagnosis of RVH in the adult population , and thus may be used as precursors to conventional catheter-based angiography. Multiple studies have evaluated the ability of ultrasonography to detect RAS in the adult population . Stenosis is diagnosed with evidence of an increase in flow velocity. In adults, a velocity >200 cm/s corresponds to >60% stenosis and an RI lower by 0.05 in the kidney behind a stenosis and/or a pulsus parvus et tardus on intrarenal spectral Doppler . Voiculescu et al demonstrated that these diagnostic criteria seem to be applicable in children and adolescents . They also suggested measurement of the flow profile in iliac arteries to help differentiate isolated RAS from obstructions that involve the aorta. Most patients in our cohort underwent angiography for definitive diagnosis of RVH. Less invasive diagnostic techniques utilized in our patient population included Doppler ultrasonography, CTA and MRA. However, false positive and false negative results are often seen utilizing these methods. Ultrasound has a sensitivity of 63% and specificity of 65% compared to a sensitivity of 88% and 80% and a specificity 81% and 63% in CTA and MRA respectively . In addition, Doppler ultrasound is poor at detecting small branch stenosis as depicted in Figure 1 . A negative intrarenal Doppler study has been shown to predict cure with intervention except in patients with Neurofibromatosis .
Location of lesions:
The location of arterial lesions in RVH varies, but has been reported to be bilateral in 53-78% of pediatric cases [5, 41, 42], with most occurring in patients who have concurrent vascular occlusions such as mid-aortic syndrome and celiac artery stenosis. In our cohort, only 24% of patients were described as having bilateral RAS. Of these, three were determined to have diffuse narrowing including other vessels (aorta, common hepatic artery, SMA) in the setting of syndromes or vasculitides. As previously described, extra- and intracranial cerebrovascular disease was also seen in our patients with RVH .
As in our patient cohort, previous studies have reported that single, focal arterial stenoses are most common in children without concomitant extra-renal vascular disease. Those with extra-renal disease or syndromes or vasculitides are more likely to have diffuse narrowing . Identification of the responsible lesion is important in management decisions, as children with intrarenal disease, bilateral renal disease and diffuse vascular disease including MAS are less likely to have improvement . Furthermore, patients with unilateral short discrete lesions (<10mm) in the main renal artery [16, 23] with <20% residual stenosis after angioplasty  are more likely to demonstrate good response in blood pressure.
The approach to management of pediatric RVH is variable and must be individualized to take into consideration each patient’s anatomy, disease etiology, and local expertise. With increasing experience in endovascular surgery, there has been a major shift in management of adult RVH. Using data from the Nationwide Inpatient Sample (NIS) database, Knipp et al showed a 73% and 56% decline in combined aortic and renal revascularizations and isolated renal revascularizations with a corresponding 173% increase in catheter-based procedures in a 13-year study period in the adult population . We saw a similar trend in our cohort which may also be attributed to increasing experience with catheter-based interventions. In our patients, the majority of patients underwent endovascular intervention (43%), with primary surgical and medication-only interventions occurring less commonly (26% and 31%, respectively).
In our experience, PTA had satisfactory outcomes in the pediatric population with 17.6% cure and 64.7% improvement rates. This is similar to other large series which ranged between 27- 39% cure and 17-40% improvement, and our re-intervention rate (24%) was also similar to other series    . This suggests that endovascular treatment is a safe option that requires careful patient selection. Stents have been infrequently utilized in certain centers with restenosis rates as high as 30%, and thus have been reserved for severe or recurrent stenosis or management of complications [17, 23].
Local experience is extremely important in determining the modality of treatment. Some centers have focused largely on open surgical interventions [11, 19, 22, 27, 47], while others have reported more experience with catheter-based interventions [17, 23, 26]. In the largest published series, open surgical interventions had high rates of cure of 70%  and 82% , compared with endovascular interventions 23% , 34%, 39% and 27% . Similarly, we saw an almost 2 times higher rate of cure with surgical intervention with about 20% of patients requiring a second surgical intervention. Of note, Eliason et al noted that remedial operations for persistent RVH after endovascular intervention were more difficult in patients who had received prior angioplasty. They also found that the risk of nephrectomy after remedial operation is significantly higher after angioplasty .
Although endovascular intervention required re-angioplasty in some patients and progression to open surgery in others, in our practice, primary surgical intervention is not typically the treatment modality of choice due to the technical challenges that exist in renal artery reconstruction and revascularization, particularly in small children with diminutive caliber vessels . The goal of surgical management is to definitively manage renovascular disease in a single operation . The decision to proceed with operation primarily was impacted by patient anatomy, in particular, location and extent of stenosis. Medication-only treatment was determined to have the least likely opportunity for cure in our population (0%), as well as the highest failure rate (25%). This is potentially due to patient factors including non-adherence, disease refractory to medical treatment, and loss to follow-up.
Although renal artery stenosis due to FMD is often successfully managed, in particular by either endovascular techniques or surgery approaches, management of RVH due to diffuse disease processes such as mid-aortic syndrome is often more complicated and carries a higher failure rate. Often associated with Takayasu arteritis or mid-aortic dysplastic syndrome, severe stenosis of the distal thoracic or proximal abdominal aorta may require aortic bypass, sometimes with adjunctive renal bypass for durability and management of hypertension . Some patients require endovascular procedures simultaneously or after surgical management. In a long term follow-up of treatment of pediatric MAS, Porras et al demonstrated that surgical intervention relieved the obstruction and resulted in a longer freedom from intervention than catheter-based interventions . Their approach employed invasive measures when medical management was insufficient and delayed definitive surgical correction until a more amenable age and size. Despite aggressive surgical interventions, residual hypertension has been reported in up to 34% of patients postoperatively . Likewise, postoperative complication rates remain high and include morbidities such as aortic rupture and dissection, bleeding, thrombosis and graft stenosis .
Limitations of our study include biases intrinsic to its retrospective design. Data available only included those which were recorded in patient medical records, and thus, additional information was unobtainable. Statistical analysis could not be performed to compare the three different treatment modalities (medication treatment only, angioplasty and open surgical approaches) as limited information was available regarding patient factors influencing choice of intervention. Additionally, current follow-up was not available on all patients and some had a short duration of follow-up. Thus, the classification of ‘cured’ may be underestimated, as patients who were requiring oral anti-hypertensive medications at the time of last follow-up may have since ceased all medical treatment. Likewise, long-term outcomes are difficult to determine. Another limitation included the 23-year span of the study. Although this time span allowed a greater number of patients to be included in the review, during this time period, innovations in techniques (both open surgical and endovascular), as well as modifications of treatment approaches, have likely influenced overall patient outcomes. A final important limitation to our study is the fact that patients with diffuse vascular disease or syndromes who often have long segment stenoses were typically treated with medication only or surgery and not considered for endovascular treatment at our institution. This may have caused a selection bias in our patient population and thus, we cannot extrapolate success of endovascular treatment in this subpopulation.
RVH in children entails a variety of heterogeneous lesions that may result in serious sequelae. Typically presenting as incidental hypertension in an asymptomatic child, the diagnostic algorithm for patients with suspected RVH includes both non-invasive imaging and conventional catheter-based angiography. Consideration should be given to the possibility of syndromes and vasculitides that may accompany RVH in children. The treatment of RVH in the pediatric population must be individualized to each patient’s anatomic characteristics, with medication-only, endovascular and surgical interventions all allowing for favorable outcomes in 79% of individuals overall. A multidisciplinary approach to the treatment of RVH should be considered including combined efforts by nephrology, interventional radiology and surgical specialties in order to achieve optimal outcomes.