Pediatr Nephrol (2002) 17:943–949 DOI 10.1007/s00467-002-0954-7
H Y P E RT E N S I O N
Chulananda D. A. Goonasekera · Vanita Shah Angie M. Wade · Michael J. Dillon
The usefulness of renal vein renin studies in hypertensive children: a 25-year experience Received: 19 April 2001 / Revised: 24 January 2002 / Accepted: 17 June 2002 / Published online: 14 September 2002 © IPNA 2002
Abstract Lateralised renin secretion predicts improvement following surgery/angioplasty of the affected kidney in paediatric hypertension. We retrospectively analysed all our renal vein renin (RVR) studies undertaken in hypertensive children and their subsequent clinical outcomes over a 25-year period. The outcomes were categorised as cure, improvement or no change in hypertension. The proportions that benefited by ‘following’ or ‘ignoring’ the RVR results in relation to the aetiology of hypertension and subsequent therapy were compared. A ratio between the RVR value of the dominant kidney and that of the contralateral kidney (R/Rc ratio) above 1.5 was considered to show significant lateralisation. A ratio between RVR value of the contralateral kidney and that of the caudal inferior vena cava (Rc/P ratio) of less than 1.3 was considered to indicate contralateral renin suppression. Of 233 RVR records, 137 were suitable for analysis. Of these, 39 had reflux nephropathy (RN), 73 renovascular disease (RVD) and the others miscellaneous conditions. Of the 39 hypertensive children with RN, the R/Rc ratio was over 1.5 in 20 subjects, of whom 11 underwent surgery. Amongst the 19 with an R/Rc less than 1.5, 4 also had surgery. Of the 73 hypertensive children with RVD, the R/Rc ratio was over 1.5 in 38 subjects, of whom 20 underwent surgery/angioplasty. Amongst the 35 with an R/Rc less than 1.5, 7 also underwent surgery/angioplasty. In RN there was no significant difference in the proportion that improved with allocated treatment, as suggested by the ‘test’ when compared with those who were ‘misallocated’. In RVD, however, C.D.A. Goonasekera · V. Shah · M.J. Dillon (✉) Department of Nephrourology, Institute of Child Health and Great Ormond Street Hospital for Children NHS Trust, 30 Guilford Street, London WC1 N1EH, UK e-mail:
[email protected] Tel.: +44-20-79052651, Fax: +44-20-79052133 A.M. Wade Department of Epidemiology and Public Health, Institute of Child Health and Great Ormond Street Hospital for Children NHS Trust, 30 Guilford Street, London WC1 N1EH, UK
the proportion improved (35.0%) by ‘following the test’ (i.e. R/Rc ratio) was significantly greater than the proportion improved (12.5%) in the group in whom the test was ‘ignored’. This was true for both R/Rc and Rc/P ratios. Thus, RVR measurements are useful for identifying hypertensive children with RVD that may have a favourable outcome after surgery, but this does not apply to children with RN. Measurements, however, may have diagnostic value even if surgery or angioplasty is not anticipated. Keywords Hypertension · Outcome · Renal vein renin · Reflux nephropathy · Renovascular disease
Introduction Renal vein renin (RVR) profiling is utilised for the evaluation of stenotic renal artery lesions in hypertensive subjects, particularly in children, due to the high prevalence of renovascular disease (RVD) in this age group (5%–25%) [1] compared with adults (1%–5%) [2, 3, 4]. Since in normotensive children without renal disease the renin secretion from each kidney is consistently symmetrical [5], this technique allows lateralisation of the kidney from which excess renin release is occurring [6] in pathological states. A ratio between the RVR value of the dominant kidney and that of the contralateral kidney (R/Rc ratio) of 1.5 has been considered a positive index for hypersecretion of renin from the affected side, although some investigators have used ratios as high as 2.5 to obtain better specificity [7, 8, 9, 10, 11, 12]. A ratio between the RVR value of the contralateral kidney and that of the caudal inferior vena cava (Rc/P ratio) of 1.5 8a 1c 2 52.2%
22 (45.8%) Yes (4) 1.5 8 a, 2 5 5
Yes (7) 1.3 2 a, 1 3 – 29.4%
21 (51.2%) No (18) 1.3 1a 7 –
No (37) 1.5 Reflux nephropathy group Renovascular disease group
Difference=24.9% Difference=22.5%
(–8.4, 58.2) (2.7, 42.3)
P=0.14 P=0.026
For Rc/P ratio 1.5 R/Rc 1.5) lateralisation of renin secretion. In RVD, however, the proportions of cases with lateralisation of renin secretion was virtually similar between the groups with unilateral (51%) and bilateral (46%) kidney involvement.
Discussion The majority of subjects who underwent RVR sampling in this study were subsequently diagnosed to have RVD, although renal scarring is the commonest cause of hypertension in children [26]. This inversion in proportions between RVD and other causes of hypertension in the study group reflects the existing clinical bias in selecting children with hypertension [23] for this investigation based on age, severity of hypertension and the assumed potential for ‘cure’ and complications. The duration of follow-up has been particularly short in some cases, as a proportion of children was transferred back to referring units, some overseas. These subjects were, therefore, excluded from the study, utilising an arbitrary cut-off of less than 6 weeks’ follow-up. The standardisation or withdrawal of antihypertensive therapy pre RVR study was not carried out in our patients due to clinical and ethical reasons. Angiotensin converting enzyme (ACE) inhibitors were not in clinical use during the early years of this study and during latter years ACE inhibitor therapy was instituted only after diagnosis or exclusion of RVD. Therefore it is unlikely that any child was on ACE inhibitor therapy at the first time of RVR study. They could have been on ACE inhibitor therapy in subsequent RVR studies, and these have been excluded from the analysis. Renin stimulation was also not attempted due to its doubtful value [27, 28]. For example, in hypertensive adults, a stimulated (diazoxide or tilt) RVR ratio is more predictive of a favourable outcome in RVD following surgery [29], whereas in renal parenchymal disease it is known to increase false-positive rates [29]. The characteristic pattern of curable renovascular hypertension in unilateral renal artery stenosis, i.e. an increased concentration of renin in the renal vein from the ischaemic kidney, together with suppression of renin secretion from the contralateral kidney [2, 30, 31], which is also seen in patients with bilateral stenoses [27], may also occur for reasons other than an absolute rise in renin secretion. Reduced or occluded blood flow in the ipsilateral kidney [27, 32], and suppression of renin secretion or net extraction of renin entering the contralateral kidney, may influence lateralisation [33], leading to an inappropriate surgical/angioplasty therapeutic approach. Furthermore, RN and RVD (the commonest causes of hypertension in children) may affect both kidneys [34], even though this is not always radiologically demonstrable. This may offer some explanation for the lack of response to surgery observed in 11.2% of patients with RN and 33.4% of patients with RVD with lateralised renin secretion in this study. However, one must also take into consideration that all surgical interventions described were
quite variable and it was unlikely that they were equally effective in treating hypertension, in particular in RVD. When considering the R/Rc ratios, the duration of follow-up was not dissimilar between ‘test-followed’ and ‘test-ignored’ groups with RVD. In the RN group, the median duration of follow-up was longer in the ‘testignored’ group (i.e. more time for spontaneous improvement and hence the probability of more cases of ‘cure’ in this group purely due to a longer duration of follow-up), but this could not have biased our results. This is because if the ‘test-followed’ group had been followed for a longer duration one could have argued that inadequate follow-up of the ‘test-ignored’ group may have biased the results in favour of the prediction of outcome by the test. The proportion of subjects who showed radiological evidence of unilateral disease was similar between ‘testfollowed’ and ‘test-ignored’ groups, and hence should not have biased the results. In our study, the finding of a significant R/Rc ratio greater than 1.5 in 30% of hypertensive children who were not diagnosed to have renovascular or parenchymal disease is difficult to explain, but has also been reported in adults [35]. It is possible that these children may have had, for instance, subtle RVD that was not recognised by the current renal imaging techniques. We are aware that some children who were initially thought to have had essential hypertension may go on to develop evidence of RVD with age. Furthermore, this group of patients had relatively lower levels of renin in their blood samples, leading to the magnitude of any errors in measurement being sufficient to influence the ratios significantly. Therefore, it is difficult to conclude that the above results indicate a ‘true’ false-positive rate. From our results, both R/Rc and Rc/P ratios were independently useful in identifying subjects who were likely to improve following surgery/angioplasty in RVD and following surgery in RN, as previously observed [14]. However, 2 of 2 patients in the RN group and 2 of 4 patients in the RVD group (Table 3) showed improvement following surgery, even in the absence of contralateral renin suppression. Furthermore, a significant proportion of subjects (≅25% in RN and ≅10% in RVD) who did not undergo surgical/angioplasty therapeutic manoeuvres also improved spontaneously with time, suggesting a ‘burn-out effect’ with age in some patients. Although the experimental evidence obtained in several animal species confirms a pathogenic role for the renin system in the early phase of induced renovascular hypertension, its role in the chronic phase of the disease is uncertain [36], suggesting a modulation in the role of renin with time. It is very difficult to investigate this non-homogeneous behaviour of the renin system in hypertensive children. This is because the duration of the disease in hypertensive children at presentation is unknown (as they could have been asymptomatic for some time before presenting) and hence the status of the renin blood pressure link (i.e. whether acutely pathogenic or chronically modulated) at diagnosis. This may also offer some explanation for the lack of response to surgery/an-
948
gioplasty observed in 1 of 3 patients with a positive R/Rc ratio in this study and the wide differences in the predictive value of the RVR ratio for cure following surgery/angioplasty observed in many previous studies. It is clear in this retrospective study that the decision to undertake surgery/angioplasty has not been totally dependent on RVR results; for example, R/Rc test results have been followed only in 26 of 41 cases in RN and in 48 of 73 cases in RVD. Only 9 subjects of 17 with RN and 18 subjects of 35 with RVD who had significant R/Rc and Rc/P ratios concomitantly were subjected to surgery or angioplasty. Therefore, in nearly 50% of cases a highly significant RVR result has been ignored. This outcome was perhaps determined by factors such as the feasibility of surgery/angioplasty (including age), parental consent, severity of hypertension, remaining kidney function, laterality of the disease and even differences in opinion amongst physicians. This has inadvertently allowed us to calculate sensitivity and specificity of the test in producing cure or improvement following surgery utilising the data. Studies in adults suggest that RVR ratios are not useful for the prediction of surgical success. A retrospective analysis of RVR results and outcome following either surgery or percutaneous transluminal angioplasty in 96 mainly adult patients with renal artery stenosis and hypertension found that almost half the patients subsequently cured or improved had an RVR ratio less than 2.0 (suggesting a poor specificity) [37]. Sellars et al. [15] demonstrated that in 37 subjects with unilateral disease the false-positive rate of the test (RVR ratio >1.5) was 39% and the false-negative rate was 71%, with little difference in outcome between those with ratios above or below 1.5 [15]. However, some other investigators, with careful inclusion criteria such as severe and refractory hypertension, unilateral renal artery stenosis with an estimated lumen obstruction to less than a quarter and exclusion criteria such as extreme age, disseminated severe atherosclerosis and reduced renal function, have shown that RVR ratios can be highly predictable of successful surgical outcome [13, 38]. Their value in the recognition of renin-secreting tumours, however, is marginal [39, 40]. The sensitivity of the R/Rc ratio in identifying patients who may benefit from surgery is high (above 80%) in RN and RVD, but the specificity is lower (50% or less) in both cases. Similarly, Rc/P ratios demonstrated a higher sensitivity in predicting cure following surgery, but specificity was extremely low (0%), in particular for RN. This suggests that lack of contralateral renin suppression should not preclude surgery/angioplasty, if otherwise indicated, in RN or RVD. Our current data generally support previous observations in our centre 20 years ago [14], where 15 children with RVD and R/Rc ratios above 1.5 were subjected to surgery (12 by nephrectomy and 3 by revascularisation) and 13 showed complete cure of hypertension (with near 100% success rate). It is noteworthy that a specificity for the test was not calculated in the previous study [14], as
none with a RVR ratio less that 1.5 underwent surgery. This explains the discrepancy between our current and previous data [14] in the context of sensitivity and specificity of the test in recognising the potential for cure following surgery. False-negative results for RVR ratio have been more commonly reported for fibromuscular disease [41], which accounts for the majority of RVD in children [42, 43, 44]. This may also have contributed to the poor specificity of our RVR results. We must, however, note that these calculations of sensitivity and specificity have been performed retrospectively, and thus reflect such indices of RVR in relation to the clinical practice in this centre, i.e. who were selected to undergo RVR studies and based on what criteria. Furthermore, it is important to recognise that our study is retrospective and there was no guarantee that had the misallocated groups been treated differently (i.e. as the tests suggested) results would have been better. Therefore, to determine whether either test would be beneficial in treatment allocation, a randomised control trial of ‘treat according to test criteria’ versus ‘treat using usual allocation criteria’ is needed. The innate value of RVR measurements, however, is not purely predictive of outcome following surgery or angioplasty, but also allows identification of the kidney or the segment of the kidney from which excess renin release is occurring. This information might be useful in determining which kidney to approach first in a surgical sense, particularly if there is bilateral disease [34], and even if surgery or angioplasty is anticipated may have diagnostic value. In conclusion, the lateralised renin secretion from one kidney, although useful in predicting outcome following surgery/angioplasty in hypertensive children with RN or RVD irrespective of unilateral or bilateral kidney involvement, should not be used on its own for determining therapy, due to its poor specificity, especially in RN. Acknowledgements We thank Ms. G. Hamill and Ms. D. Doyle for their invaluable assistance in tracing medical records. We are grateful to all our colleagues in the Departments of Paediatric Cardiology and Paediatric Radiology who have assisted us over the years in performing these invasive investigations in hypertensive children. The project was funded by John Herring and Friends Fund (Child Health Research Appeal Trust).
References 1. Watson AR, Balfe JW, Hardy BE (1985) Renovascular hypertension in childhood: a changing perspective in management. J Pediatr 106:366–372 2. Pickering TG, Sos TA, Vaughan ED Jr, Case DB, Sealey JE, Harshfield GA, Laragh JH (1984) Predictive value and changes of renin secretion in hypertensive patients with unilateral renovascular disease undergoing successful renal angioplasty. Am J Med 76:398–404 3. Pickering TG, Sos TA, Vaughan ED Jr, Laragh JH (1986) Differing patterns of renal vein renin secretion in patients with renovascular hypertension, and their role in predicting the response to angioplasty. Nephron 44[Suppl 1]:8–11 4. Hillman BJ (1989) Imaging advances in the diagnosis of renovascular hypertension. Am J Roentgenol 153:5–14
949 5. Gerdts KG, Shah V, Savage JM, Dillon MJ (1979) Renal vein renin measurements in normotensive children. J Pediatr 95: 953–958 6. Godard A (1977) Predictive value of renal-vein renin measurements in children with various forms of renal hypertension. An international study. Helv Paediatr Acta 32:49–57 7. Amsterdam EA, Couch NP, Christlieb AR, Harrison JH, Crane C, Dobrzinsky SJ, Hickler RB (1969) Renal vein renin activity in the prognosis of surgery for renovascular hypertension. Am J Med 47:860–868 8. Stockigt JR, Collins RD, Noakes CA, Schambelan M, Biglieri EG (1972) Renal-vein renin in various forms of renal hypertension. Lancet I:1194–1197 9. Bourgoignie J, Kurz S, Catanzaro FJ, Serirat P, Perry HM Jr (1970) Renal venous renin in hypertension. Am J Med 48: 332–342 10. Figueroa JE, Bennett DJ, DeCamp PT, Batson HM (1975) Experience with renal vein renin ratios in the identification of a pressor kidney. South Med J 68:1200–1206 11. Ernst CB, Bookstein JJ, Montie J, Baumgartel E, Hoobler SW, Fry WJ (1972) Renal vein renin ratios and collateral vessels in renovascular hypertension. Arch Surg 104:496–502 12. Sonkodi S, Abraham G, Mohacsi G (1990) Effects of nephrectomy on hypertension, renin activity and total renal function in patients with chronic renal artery occlusion. J Hum Hypertens 4:277–279 13. Jensen G, Zachrisson BF, Delin K, Volkmann R, Aurell M (1995) Treatment of renovascular hypertension: one year results of renal angioplasty. Kidney Int 48:1936–1945 14. Dillon MJ, Shah V, Barratt TM (1978) Renal vein renin measurements in children with hypertension. BMJ 2:168–170 15. Sellars L, Shore AC, Wilkinson R (1985) Renal vein renin studies in renovascular hypertension – do they really help? J Hypertens 3:177–181 16. Roubidoux MA, Dunnick NR, Klotman PE, Newman GE, Cohan RH, Kadir S, Svetkey LP (1991) Renal vein renins: inability to predict response to revascularization in patients with hypertension. Radiology 178:819–822 17. Rossi GP, Pavan E, Chiesura-Corona M, Bader M, Paganini G, Cesari M, De Toni R, Feltrin GP, Ganten D, Pessina AC (1997) Renovascular hypertension with low-to-normal plasma renin: clinical and angiographic features. Clin Sci (Lond) 93:435–443 18. Young DF, Cholvin NR, Roth AC (1975) Pressure drop across artificially induced stenoses in the femoral arteries of dogs. Circ Res 36:735–743 19. Young DF, Cholvin NR, Kirkeeide RL, Roth AC (1977) Hemodynamics of arterial stenoses at elevated flow rates. Circ Res 41:99–107 20. Sealey JE, Laragh JH (1973) Searching out low renin patients: limitations of some commonly used methods. Am J Med 55:303–314 21. Maxwell MH, Varady PD (1976) Cooperatives study of renovascular hypertension. Clinical characteristics, diagnostic tests and results of surgery. Contrib Nephrol 3:1–30 22. Holley KE, Hunt JC, Brown AL, Kinkaid OW, Sheps SG (1964) Renal artery stenosis. Am J Med 37:14 23. Gordon I (1994) Imaging in systemic hypertension in paediatrics. J Hum Hypertens 8:377–379 24. Dillon MJ, Ryness JM (1975) Plasma renin activity and aldosterone concentration in children. BMJ 4:316–319 25. Dillon MJ (1975) Measurement of plasma renin activity by semi-micro radio immunoassay of generated angiotensin I. J Clin Pathol 28:625–630
26. Still JL, Cottom D (1967) Severe hypertension in childhood. Arch Dis Child 42:34–39 27. Pickering TG, Sos TA, James GD, Vaughan ED Jr, Sealey JE, Laragh JH (1985) Comparison of renal vein renin activity in hypertensive patients with stenosis of one or both renal arteries. J Hypertens 3:S291–S293 28. Kutkuhn B, Godehardt E, Kunert J, Torsello G, Grabensee B (1996) Acute blockage of the renin system and differential renal vein renin determinations in the diagnosis of renovascular hypertension. Scand J Urol Nephrol 30:69–72 29. Gordon RD, Tunny TJ, Evans EB, Fisher PM, Jackson RV (1986) Unstimulated renal venous renin ratio predicts improvement in hypertension following nephrectomy for unilateral renal disease. Nephron 44[Suppl 1]:25–28 30. Vaughan ED Jr, Buhler FR, Laragh JH, Sealey JE, Baer L, Bard RH (1973) Renovascular hypertension: renin measurements to indicate hypersecretion and contralateral suppression, estimate renal plasma flow, and score for surgical curability. Am J Med 55:402–414 31. Vaughan ED Jr, Case DB, Pickering TG, Sosa RE, Sos TA, Laragh JH (1984) Clinical evaluation of renovascular hypertension and therapeutic decisions. Urol Clin North Am 11: 393–407 32. Webb DJ, Cumming AM, Adams FC, Hodsman GP, Leckie BJ, Lever AF, Morton JJ, Murray GD, Robertson JI (1984) Changes in active and inactive renin and in angiotensin II across the kidney in essential hypertension and renal artery stenosis. J Hypertens 2:605–614 33. Malatino LS, Cumming AM, Hodsman GP, Leckie BJ, Lever AF, Tillman DM, Webb DJ, Morton JJ, Robertson JI (1986) Factors affecting renal vein renin ratio in renal artery stenosis. Secretion of inactive renin. Nephron 44 [Suppl 1]:68–72 34. Deal JE, Snell ME, Barratt TM, Dillon MJ (1992) Renovascular disease in childhood. J Pediatr 121:378–384 35. Carmichael DJ, Mathias CJ, Snell ME, Peart S (1986) Detection and investigation of renal artery stenosis. Lancet I:667–670 36. Brown JJ, Davies DL, Morton JJ, Robertson JI, Cuesta V, Lever AF, Padfield PL, Trust P (1976) Mechanism of renal hypertension. Lancet I:1219–1221 37. Rappelli A, Glorioso N, Madeddu P, Dessi-Fulgheri P, Leoni C, Fiori C, Manca GB, Sanna G, Arzilli F, Pedrinelli R, et al (1986) Renal vein renin in renovascular hypertension: the experience of two Italian centers. Nephron 44[Suppl 1]:12–16 38. Pedersen EB, Danielsen H, Fjeldborg O, Kornerup HJ, Madsen B (1986) Renovascular hypertension. Ability of renal vein ratio to predict the blood pressure level 18–24 months after surgery. Nephron 44[Suppl 1]:29–31 39. McVicar M, Carman C, Chandra M, Abbi RJ, Teichberg S, Kahn E (1993) Hypertension secondary to renin-secreting juxtaglomerular cell tumor: case report and review of 38 cases. Pediatr Nephrol 7:404–412 40. Baert J, Van Damme B, Oyen R, Van Poppel H, Baert L (1995) Juxtaglomerular cell tumor: importance of clinical suspicion. Urol Int 54:171–174 41. Marks LS, Maxwell MH, Varady PD, Lupu AN, Kaufman JJ (1976) Renovascular hypertension: does the renal vein renin ratio predict operative results? J Urol 115:365–368 42. Fry WJ, Ernst CB, Stanley JC, Brink B (1973) Renovascular hypertension in the pediatric patient. Arch Surg 107:692–698 43. Lawson JD, Boerth R, Foster JH, Dean RH (1977) Diagnosis and management of renovascular hypertension in children. Arch Surg 112:1307–1316 44. Wise KL, McCann RL, Dunnick NR, Paulson DF (1988) Renovascular hypertension. J Urol 140:911–924
