Follow-Up on 100 Renal Vein Renin Samplings

Journal of Human Hypertension (2002) 16, 275–280  2002 Nature Publishing Group All rights reserved 0950-9240/02 $25.00 www.nature.com/jhh ORIGINAL ARTICLE Hypertension and renovascular disease: follow-up on 100 renal vein renin samplings

P Hasbak1, LT Jensen2, H Ibsen3, and The East Danish Study Group on Renovascular Hypertension* 1Department of Clinical Physiology and Nuclear Medicine, Glostrup Hospital, Denmark; 2Department of Clinical Physiology and Nuclear Medicine, Gentofte Hospital, Denmark; 3Department of Internal Medicine, Glostrup Hospital, Denmark

The clinical value of renal vein renin sampling (RVRS) р15%). None of the indices clearly discriminated as a prognostic tool in the treatment of renovascular between the patients who did benefit from intervention, hypertension was evaluated. One hundred consecutive and those who did not. The only positive finding was patients were included over a 4-year period of time. that a peripheral renin concentration lower than 8 mlU/l About half of the patients (49%) were treated inter- predicted no effect of intervention, which might lead to ventionally by PTRA (21%), nephrectomy (20%), or vas- the exclusion of 11% of the patients before entering the cular surgery (8%). Seven patients (15%) were cured and diagnostic programme. We conclude that the RVRS 15 (32%) had improved (reduction in antihypertensive demands a very restrictive referral pattern if it should medicine) after 6 months follow-up, whereas three be of prognostic value for the blood pressure outcome patients (6%) were cured and 12 (26%) improved after after intervention. No indices of lateralised renin con- 3–4 years follow-up. Thus, the number of patients cured centrations proved high predictive value. However, a or improved is comparable with the results from our peripheral renin concentration low in the normal range department reported 20 years ago. However, in the seems useful as an indicator of no benefit from inter- present report, more than twice as many patients were vention. enrolled, leading to double costs. Different indices of Journal of Human Hypertension (2002) 16, 275–280. DOI: lateralisation of the renin generation were calculated for 10.1038/sj/jhh/1001365 the use in cases of a shrunken kidney (functional share

Keywords: renovascular hypertension; renal vein renin sampling; cure and improvement rate

Introduction tors, in addition to ␤-adrenoreceptor blockers, have made the pharmacological management of arterial More than 20 years ago, the results of interventional hypertension a possible competitor to interventional treatment of renovascular hypertension in our 2,3 1 treatment of renovascular hypertension. Further- department were published. At that time, nephrec- more, the prevalence of renovascular hypertension tomy (82%) and vascular surgery (18%) were the is now estimated to be less than 0.5%4 of the hyper- treatments of choice. Since then, the percutaneous tensive population, in contrast to earlier estimates transarterial renal angioplasty (PTRA) has become up to 5–10%.2 The therapeutic changes combined the preferred treatment of renal artery stenosis. In with the conflicting reports on the clinical useful- the same period of time, the antihypertensive drugs ness of renal vein renin sampling (RVRS) and on the have become increasingly effective. The introduc- long term effects of PTRA on hypertension,5–10 tion of calcium channel blockers and ACE inhibi- recently addressed by van Jaarsveld and col- leagues,11 prompted us to evaluate our strategy in Correspondence: P Hasbak, MD, Department of Clinical Physi- diagnosing and predicting the outcome of inter- ology and Nuclear Medicine, Glostrup Hospital, DK-2600 ventional treatment of renovascular hypertension. Glostrup, Denmark. E-mail: philipȰpost1.tele.dk The study was retrospectively made on 100 con- *Members of the study group: H Dige-Petersen, A Leth, M Brahm, secutive patients who underwent RVRS. The aim of S Schifter, S Rasmussen, Glostrup Hospital; S Strandgaard, S Dorph, F Rasmussen, Herlev Hospital; JK Christoffersen, JO Lund, the study was to investigate if optimisation of indi- S Just, N Baekgaard, LP Jensen Gentofte Hospital; A Høegholm, ces calculated from the RVRS data would improve Naestved Hospital; K Rasmussen, Roskilde Hospital. the prognostic power to predict the outcome of Received 7 August 2001, revised and accepted 29 October 2001 PTRA or nephrectomy on arterial hypertension, and Follow-up on 100 renal vein renin samplings P Hasbak et al 276 to estimate the cost of curing renovascular hyperten- radiologists, who also performed the PTRA pro- sion. cedures, interpreted all angiographies. The RVRS was successful in all patients. Treat- ␣ ␤ Subjects and methods ment with -/ -adrenoreceptor blockers, ACE-I and angiotensin 2 receptor (AT2) antagonists was with- The study population consisted of 100 patients con- drawn for at least 2 weeks prior to RVRS. Owing to secutively referred to Glostrup County Hospital for a thrombotic inferior caval vein, the catheterisation RVRS from 1994 to 1997. For basic data, see Table procedure was performed from the right cubital vein 1. The study was initiated in April 1998, and with in two patients. The remaining patients all had two a late follow-up in 2000. The reference population catheters, from the left and right femoral vein to the was 1.3 million people in the East of Denmark, left and right renal vein. The correct position of the excluding the city of Copenhagen. Approximately catheters was ensured by fluoroscopy, measurement 700000 people were older than 40 years. Thirteen of oxygen tension, and by measuring 51Cr-EDTA in percent (91000 people) received antihypertensive the renal veins in comparison with the reference drugs, with 0.5% of the population (3500 people) sample (iliac vein blood). At start of the procedure, receiving three or more different drugs.12 51Cr-EDTA (3.7 MBq) was given intravenously. Sim- ultaneously blood samples were taken before and after stimulation with furosemide. Plasma renin Diagnostic procedures concentration was determined using the principle of The diagnosis of renovascular hypertension was antibody trapping,14 as modified by Millar et al.15 based on the history of severe hypertension (three The standard renin preparation, 68/356, was or more different antihypertensive drugs/rapid onset obtained from the National Institute for Biological of hypertension), renography, renography with Standards and Control (Hertfordshire, UK) and anti- acute angiotensin-converting enzyme-inhibition bodies against angiotensin I were raised in rabbits. (ACE-I), renal artery angiography and renal vein The decision for a ‘positive RVRS’ was based on a renin sampling. Usually renal artery angiography significant renin gradient across the ipsilateral kid- was carried out before RVRS except in cases where ney and significant suppression of renin secretion the referring physician believed that the outcome of on the contralateral site. An expert panel of phys- RVRS could lead to a decision as far as intervention icians and surgeons made the decision about inter- was concerned. vention after presentation of all clinical and diag- The renographies (without and with acute ACE-I) nostic data. were carried out according to the international rec- ommendations,13 and interpreted by the same panel Follow-up and evaluation of four experts in nuclear medicine. The renal angiographies were made with a right After PTRA, vascular surgery or nephrectomy the and a left view of 15 degrees from centre. In two patients were followed at their local hospital for patients the angiography was not technically poss- blood pressure controls and evaluation. To deter- ible due to extensive atherosclerosis, and in one mine possible benefit of the intervention, the patient patient the angiography was performed via trans- files were reviewed regarding post treatment blood lumbal catheterisation of the aorta, owing to pressure and antihypertensive therapy. If the infor- occlusion of both femoral arteries. The remaining mation was available the blood pressure is given as angiographies were performed with aortic catheris- the mean value of three consultations around the ation from the right femoral artery. The same two time of intervention or RVRS, at 6 months and 3–7 years after intervention/RVRS. A patient was considered ‘cured’ if he/she became normotensive (diastolic blood pressure р90 mm Hg) Table 1 Basic characteristics of the study population in the absence of any antihypertensive medication for at least 6 months. Improvement was defined as Number of patients included 100 Patients excluded 6a a reduction in the need for antihypertensive medi- Study population 94 cation. If none of these conditions were fulfilled the Sex (Female/Male) 37/57 patient’s hypertension was considered ‘unchanged’. Age (years) 58.6 (33.4–84.9)b Blood pressure (mm Hg) Systolic 170 (125–240)b Database, calculations and statistical analysis Diastolic 95 (64–135)b S-creatinine (␮mol/l) 102 (60–536)b Data were collected in a base containing the follow- Antihypertensive medicine ing on each patient: age, gender, serum creatinine one drug 21 (prior to intervention), renography data (functional two drugs 29 three or more drugs 44 share/shape of renograms), renin values from the RVRS (basal conditions and after stimulation with aFour patients died within 6 months, two settled in other areas furosemide), antihypertensive medication (at time of of Denmark. bData are given as median (range). RVRS and at follow up), result of renal angiography,

Journal of Human Hypertension Follow-up on 100 renal vein renin samplings P Hasbak et al 277 interventional procedure (PTRA, nephrectomy, vas- was whether a nephrectomy should be performed. cular surgery), and blood pressure (mean of three In all patients, arteriosclerosis was the cause of the consultations before intervention, and at follow up). stenosis, and none due to fibromuscular dysplasia. Four different indices to quantify the lateralis- Interventional treatment was applied in 47 patients ation of the renal renin out-put were calculated, (Figure 1, Table 2) and the result on blood pressure attempting to optimise the prognostic value of the was seven cured, 15 improved and 25 unchanged RVRS (Figure 1). One of the indices was the ‘Renin- after 6 months follow-up, whereas only three Generation-Index’, estimating the net renin release remained cured and 12 improved at the 3–4 years from each kidney by correcting the net renal vein follow-up (Figure 1). In 27 patients with angiograph- renin production (difference between renal vein and ically verified stenoses, intervention was omitted reference) with the renographically determined either because the RVRS was considered negative or functional share (Figure 1).16 the degree of renal artery stenosis was not sufficient The Mann–Whitney U-test was used to test differ- for PTRA, the complexity of the patient or refusal ences between subgroups. P-values less than 0.05 to participate. were considered statistically significant. Data are The indices for lateralisation of renin secretion, expressed as medians and ranges. given in Figure 2, were calculated, and different cut off levels tested. However, none of them showed to Results be more valid for the prediction of blood pressure outcome after interventional treatment, than the One hundred patients entered the study and underwent RVRS. Baseline characteristics are presented in Table 1. Four patients died before the Table 2 Intervention on 47 patients and number of patients cured first follow-up, and two settled in other parts of and improved after 6 months follow-up Denmark. The remaining 94 patients all had an Intervention Total Cured Improved abnormal renography/ACE-I-renography (Figure 1). (6 months) (6 months) In 60 patients their RVRS was considered ‘positive’ and in 34 ‘negative’. In 85 patients renal angiogra- PTRA (R/L) 20 (12/8) 4 8 phy were performed. Seventy-two patients had renal Nephrectomy (R/L) 19 (5/14a)2 5 artery stenosis (46 unilateral and 26 bilateral), diag- Vascular surgery 9 (2a/1/6) 1 2 nosed by selective renal angiography (Figure 1). Of (R/L/Bilat.) the remaining 22 subjects, 13 had normal renal angi- PTRA, percutaneous transarterial renal angioplasty; R, right ography and in nine patients the renal angiography sided; L, left sided; Bilat., bilateral intervention. aOne patient was omitted, owing to one shrunken kidney and underwent both left sided nefrectomy and vascular surgery on normal contralateral renogram, where the question the right sided renal artery.

Figure 1 Flow-chart illustrating the 100 patients in the study population undergoing renography, renal vein renin sampling, angio- graphy and intervention. Numbers in boxes represent number of patients.

Journal of Human Hypertension Follow-up on 100 renal vein renin samplings P Hasbak et al 278 Discussion

The renal vein renin sampling procedure (RVRS) has for several decades been one of the main diag- nostic tests for renovascular hypertension. It has been argued that although the diagnostic value is good,4,18 the prognostic value of the test varies widely between centres.3,5,7 Recent reviews do not Figure 2 Renal vein renin indices are calculated using the renin even mention the RVRS as a method of evaluating concentration of the renal vein with the highest renin concen- patients suspected of renovascular hypertension.2,19 tration (Abnormal), the renin concentration from the contralateral Throughout the years our group has been a strong renal vein (Normal), the renin concentration from the infrarenal vena cava (Reference) and renograms without angiotensin- proponent for RVRS and it has a strong position in converting enzyme inhibition determined the functional Share of our decision-making. However, it should be stressed each kidney. that a proper comprehensive study to define the pre- dictive value of a positive versus negative RVRS has so far not been carried out. often used lateralisation index of 1.5 (Figure 3).1,17 The results of 100 renal vein renin samplings from The peripheral renin, used as reference value in the our department show a low predictive value for the RVRS, showed to be the most effective parameter. outcome of interventional procedures. The rate of All patients considered cured or improved had per- cure (15%) and improvement (32%) of 47 treated ipheral renin concentrations above 8 mlU/l (normal patients after 6 months follow-up is significantly range 6–60 mlU/l). lower than it was 20 years ago. However, the number The estimated cost of 4000 renographies of 100 of patients benefiting from intervention is US$, 1200 ACE-I-renographies of 120 US$, 250 unchanged about six per year per one million citi- angiographies of 600 US$, 100 renal vein renin sam- zens. The characteristics of the patients does not plings of 720 US$, and 47 PTRA/surgical inter- seem to have changed considerably; we still diag- vention of 2400 US$ in our patient population was nose very few or none with fibromuscular dysplasia in total about 720000 US$. Hence, the net cost of or arteriitis. The predominant underlying disease each cured patient was 90000 US$, or 30000 US$ continues to be atherosclerosis. The only clear-cut for each cured or improved. These costs are clearly difference is the increasing number of patients sus- underestimated. If the total costs should be esti- pected of renovascular hypertension referred for mated (buildings, diagnostic utensils, heat, elec- RVRS. In the mid-nineties we performed more than tricity etc) the cost could be multiplied with at least twice as many renal vein renin samplings as in the a factor of two. If so, the cost is still lower, but com- seventies. It is tempting to conclude that the more parable with those of Blaufox et al2 in 1996. The liberal access to renography, angiography and renal cost for the society in lost working hours because of vein renin sampling, has led to a less critical referral hospitalisation was not estimated. pattern, with an unchanged population benefiting

Figure 3 Renal vein renin values for each of the 94 patients in the study population are plotted using the lateralisation and suppression indices defined in Figure 2.

Journal of Human Hypertension Follow-up on 100 renal vein renin samplings P Hasbak et al 279 from the tests, but for the inconvenience of more programme at all, diminishing the costs signifi- patients. The consequence of referring the double cantly. However, due to the retrospective character number of patients seems just to be the double cost of our study, the finding has to be tested prospec- for each cured patient. In the present study the esti- tively. mated net cost of one cured patient is at least It is worth noting that up to 10% of all renal artery 90000 US$. stenoses will occlude within 2 years and the renal The treatment of hypertension with renovascular function will deteriorate in about 30%.19,20 The pres- disease has dramatically changed. Twenty years ago ervation of kidney function by PTRA is to our the treatment in most cases was nephrectomy and knowledge not investigated in detail despite it being in a few patients surgical angioplasty, with a cure a highly important issue. It is known that the PTRA and improvement rate of 86%.1 Our rate of cure and does not prolong the life expectancy, but perhaps improvement are of the same magnitude as reported the quality of life.23 in several other studies2,7,9,20 but much lower than reported from the centres with the highest suc- Conclusion cess.5,6,21 Several factors might explain this differ- ence, but especially the clinical selection for the first In conclusion, the same number of patients benefit renography seems important. It has been calculated from PTRA as previously described 20 years ago. that the probability of renovascular hypertension The referral pattern has changed, so now too many should be higher than 30% to obtain reasonable cost are suspected for renovascular hypertension. The benefit of the diagnostic programme.2,19 It is tempt- centres with high curerates and high benefit of renal ing to speculate that not only structural vascular vein renin sampling seem to be the most restrictive. changes in the contralateral kidney might determine For the optimal use of RVRS, we advocate a more the outcome of PTRA, ie restoring the renal artery restrictive referral pattern and the need for multi- blood flow by PTRA might not result in normoten- centre clinical databases. Especially, there is a need sion because of structural vascular changes in for studies elucidating the effect of PTRA on renal smaller arteries distal to the renal artery stenosis, function. There is an urgent need for measures of a perpetuating the hypertension. This might explain higher predictive value than RVRS such as the renal why a high renal resistance index at the stenosis resistance index. side seems to be a highly reliable predictor for non- 22 improvement after PTRA. References The most positive report concerning RVRS within the recent years6 shows that the RVRS is the test 1 McNair A et al. A follow-up study of hypertensive with the highest prognostic value. This certainly is patients after operative treatment of unilateral renovas- based on a much more restrictive referral pattern, cular or renal disease. Acta Med Scand 1979; 205: resulting in the diagnosis and treatment of approxi- 569–574. mately one patient per 100000 citizens, whereas we 2 Blaufox MD, Middleton ML, Bongiovanni J, Davis BR. Cost efficacy of the diagnosis and therapy of renovas- have applied the programme on three times as many cular hypertension. J Nucl Med 1996; 37: 171–177. patients, with the same result, approximately one 3 Derkx FH et al. Renal artery stenosis towards the year cured or improved per 100000. The difference 2000. J Hypertens Suppl 1996; 14: S167–S172. between one and three patients investigated is 4 Mann SJ, Pickering TG. Detection of renovascular tremendous, looking at the inclusion investigation: hypertension. State of the art: 1992 Ann Intern Med instead of 4000 renographies, 1200 ACE-I-reno- 1992; 117: 845–853. graphies, and 250 angiographies, we should only 5 Bergrem H, Jervell J, Solheim DM, Flatmark A. Prog- make one-third. So it may very well be possible that nostic value of renal vein renin determination in sus- 0.5% of all hypertension are renovascular, but also pected renovascular hypertension. Acta Med Scand that only about 0.1% are curable. 1982; 211: 387–391. 6 Jensen G et al. Treatment of renovascular hyperten- Our attempt to optimise the prognostic value of sion: one year results of renal angioplasty. Kidney Int RVRS by calculating more indices of lateralised 1995; 48: 1936–1945. renin release was not successful. Even the assump- 7 Kuhlmann U et al. Long-term experience in percu- tion that the ‘Renin-Generation-Index’ would be taneous transluminal dilatation of renal artery sten- valuable when dealing with shrunken kidneys was osis. Am J Med 1985; 79: 692–698. not correct. There was no difference between 8 Roubidoux MA et al. Renal vein renins: inability to patients with a shrunken kidney and patients with predict response to revascularization in patients with just moderately diminished functional share of one hypertension. Radiology 1991; 178: 819–822. of the kidneys. The only positive finding was that 9 Sellars L, Shore AC, Wilkinson R. Renal vein renin peripheral renin concentrations of less than 8 mlU/l studies in renovascular hypertension – do they really help? J Hypertens 1985; 3: 177–181. were not found in any patient cured or improved 10 Svetkey LP et al. Prospective analysis of strategies for after intervention. Using this cut-off level, 10 (11%) diagnosing renovascular hypertension. Hypertension of the 94 patients could be spared for the total inves- 1989; 14: 247–257. tigatory programme. If correct, even more patients 11 van Jaarsveld BC et al. The effect of balloon angio- could be excluded before entering the diagnostic plasty on hypertension in atherosclerotic renal-artery

Journal of Human Hypertension Follow-up on 100 renal vein renin samplings P Hasbak et al 280 stenosis. Dutch Renal Artery Stenosis Intervention teral suppression, estimate renal plasma flow, and Cooperative Study Group. N Engl J Med 2000; 342: score for surgical curability. Am J Med 1973; 55: 1007–1014. 402–414. 12 Rasmussen S. Unpublished data from the Glostrup 18 Pickering TG, Sos TA, Vaughan EDJ, Laragh JH. Dif- Population Survey: MONI 10 study. Unpublished data fering patterns of renal vein renin secretion in patients from the Glostrup Population Survey: MONI 10 with renovascular hypertension, and their role in pre- study 1998. dicting the response to angioplasty. Nephron 1986; 44 13 Taylor A et al. Consensus report on ACE inhibitor (Suppl 1): 8–11. renography for detecting renovascular hypertension. 19 Radermacher J, Brunkhorst R. Diagnosis and treatment Radionuclides in Nephrourology Group. Consensus of renovascular stenosis – a cost-benefit analysis. Group on ACEI Renography. J Nucl Med 1996; 37: Nephrol Dial Transplant 1998; 13: 2761–2767. 1876–1882. 20 Morganti A. Renal angioplasty: better for treating 14 Poulsen K, Jorgensen J. An easy radioimmunological hypertension or for rescuing renal function? J Hyper- microassay of renin activity, concentration and sub- tens. 1999; 17: 1659–1665. strate in human and animal plasma and tissues based 21 Pedersen EB et al. Renovascular hypertension. Ability on angiotensin I trapping by antibody. J Clin Endo- to renal vein ratio to predict the blood pressure level crinol Metab 1974; 39: 816–825. 18–24 months after surgery. Nephron 1986; 44 (Suppl 15 Millar JA et al. A microassay for active and total renin 1): 29–31. concentration in human plasma based on antibody 22 Radermacher J et al. Use of Doppler ultrasonography trapping. Clin Chim Acta 1980; 101:5–15. to predict the outcome of therapy for renal-artery 16 Sato K et al. Renal vein plasma renin activity in stenosis. N Engl J Med 2001; 344: 410–417. patients, with unilateral renovascular hypertension. 23 Isles C et al. Survival associated with renovascular dis- Jpn Circ J 1988; 52: 431–436. ease in Glasgow and Newcastle: a collaborative study. 17 Vaughan ED et al. Renovascular hypertension: renin Scott Med J 1990; 35:70–73. measurements to indicate hypersecretion and contrala-

Journal of Human Hypertension