Author + information
- Received September 10, 2001
- Revision received November 7, 2001
- Accepted December 17, 2001
- Published online March 6, 2002.
- ↵*Reprint requests and correspondence: Prof. A. D. Struthers, University Department of Clinical Pharmacology and Therapeutics, Ninewells Hospital and Medical School, Dundee DD1 9SY, United Kingdom
This study was designed to fully characterize vascular tissue angiotensin I (AI)/angiotensin II (AII) conversion changes over time in vivo in humans during chronic angiotensin-converting enzyme (ACE) inhibitor therapy.
Plasma AII does not remain fully suppressed during chronic ACE inhibitor therapy. However, the plasma renin angiotensin system (RAS) might be dissociated from the vascular tissue RAS. We therefore set out to characterize the time course of vascular RAS reactivation during chronic ACE inhibitor therapy.
Vascular AI/AII conversion was studied in patients with chronic heart failure (CHF) taking chronic lisinopril therapy by the differential infusion of AI and AII into the brachial artery. A cross-sectional study was done to see whether there were differences in vascular AI/AII conversion according to New York Heart Association (NYHA) class. A second longitudinal study followed 28 patients with NYHA I to II CHF serially over 18 months to see whether vascular ACE inhibition was progressively lost with time despite ACE inhibitor therapy. A third study examined whether increasing the dose of lisinopril affected subsequent vascular ACE inhibition.
In the cross-sectional study, vascular AI-to-AII conversion was significantly reduced in NYHA class III compared with class I/II (p < 0.05). In the longitudinal study, vascular ACE inhibition was significantly reduced at 18 months as compared with baseline (p < 0.001), suggesting gradual reactivation of vascular ACE in CHF over time. In the third study, tissue ACE inhibition could be restored by increasing the ACE inhibitor dose.
Vascular AI/AII conversion reactivates over time during chronic ACE inhibitor therapy even if the CHF disease process is clinically stable. It also occurs as the CHF disease process progresses. Even if vascular AI/AII conversion has reactivated, it can be suppressed by increasing the dose of the ACE inhibitor.
Angiotensin-converting enzyme (ACE) inhibitors have clearly been a major advance in the treatment of chronic heart failure (CHF), but cross-sectional studies have suggested that plasma angiotensin II (AII) does not remain suppressed in all patients with CHF during chronic ACE inhibitor therapy (1,2). However, plasma levels of AII are not that informative because it is now well accepted that there are two very different components to the renin angiotensin system (RAS); there is the circulating RAS and the tissue RAS (3–5), and several indirect lines of evidence suggest that the circulating RAS and the tissue RAS are often dissociated (5–7).
The question naturally arises whether vascular tissue ACE reactivates over time during chronic ACE inhibitor therapy. Preliminary evidence, again only from cross-sectional studies, shows that human tissues can sometimes generate AII from angiotensin I (AI) during chronic ACE inhibitor therapy (8,9). However, cross-sectional studies where different individuals are studied only once are clearly an imperfect, if not a flawed, way to study a phenomenon thought to occur gradually within an individual over time. The definitive way to study such a phenomenon is to do a longitudinal study in which the same individuals are studied repeatedly. But there are as yet no longitudinal studies to examine whether vascular tissue AI/AII conversion reactivates progressively over time within an individual during chronic ACE inhibitor therapy. There are also no data to see whether vascular AI/AII conversion is reactivated even more when CHF disease progresses as it usually does. We sought to answer both these questions, and also considered whether increasing the ACE inhibitor dose would help when vascular AI/AII conversion has reactivated. After all, it is possible that during chronic ACE inhibitor therapy, non-ACE pathways could contribute to vascular tissue AII generation, such that increasing the ACE inhibitor dose further at the late stages of the disease would have little impact on vascular tissue AII generation. A fourth question that we addressed was whether reactivation of vascular tissue AII generation was dissociated from reactivation of the plasma RAS.
Study populations (table 1)
All studies received local ethical approval and all subjects gave written informed consent.
Study 1: cross-sectional study
In the first study, 37 patients (age 64 ± 7 years) with New York Heart Association (NYHA) class I to III CHF were studied in a cross-sectional fashion using a standard vascular function protocol to see whether any differences in vascular ACE activity could be observed according to disease severity by NYHA class. The three groups (NYHA I, II and III) were matched by age, gender and ACE inhibitor dose. All patients were taking the ACE inhibitor lisinopril to avoid any confounding issues with regard to differences in tissue ACE binding properties with different ACE inhibitors. Unfortunately, NYHA class IV patients cannot be studied with this technique because they cannot lie flat for several hours.
Study 2: longitudinal study
Twenty-eight patients with NYHA class I and II CHF taking a fixed dose of lisinopril were studied at baseline, after nine months and after 18 months to see whether there was any progressive loss of vascular ACE inhibition over time despite chronic ACE inhibitor therapy.
Study 3: changing ACE inhibitor dose study
Ten patients with the least vascular ACE inhibition seen in Study 1 were restudied on three further occasions, taking a baseline dose of lisinopril of 10 mg/day. Subjects had their vascular function assessed at baseline, after taking lisinopril 5 mg/day for six weeks and after taking 20 mg/day for six weeks. This was a double-blind crossover study with randomization of treatment allocations. The study was designed to see whether vascular AI/AII conversion could be suppressed by increasing the dose of the ACE inhibitor.
Vascular function protocol
We assessed vascular AI/AII conversion by forearm venous occlusion plethysmography using a protocol described in detail by our group (10,11). All study visits were carried out in the afternoon (2 pm), 6 h after dosing with lisinopril to ensure peak drug effect.
First, AI was infused at 16, 64, 256, 512 and 1,024 pmol/min, each for 7 min, to produce a cumulative dose-response curve. This was followed by AII infused at 4, 16, 64, 256 and 1,024 pmol/min for 7 min each. Between the AI infusion and the AII infusion, the infusion set was flushed with saline, and sufficient time was allowed for the forearm blood flow to return to baseline values (approximately 20 to 30 min).
Angiotensin I only exerts its vasoconstrictive effect in this forearm model through conversion in the vasculature to AII, and therefore the vasoconstriction elicited by AI reflects vascular AI/AII conversion (12,13). This technique has been validated and is now used extensively by us and many other investigators (12,13).
For all studies, markers of plasma RAS activity (AI and II, aldosterone, plasma ACE) were analyzed. Blood sampling was carried out using intravenous cannulae (20 G) placed into dorsal hand veins to cause a minimum of discomfort and disturbance to resting subjects. Once blood samples were obtained, aliquots were centrifuged at 4°C, separated and stored at −70°C (AI and II) or −20°C (aldosterone, ACE) until assayed simultaneously in a single batch.
Plasma AI and AII
Blood samples were taken into chilled tubes containing ethylene diamine tetraacetic acid, enalkiren (a renin inhibitor), enalapril (an in vitro ACE inhibitor) and O-phenanthroline. In the presence of an ACE inhibitor drug, chronic ACE inhibition prevents ACE-mediated generation of AII, and feedback suppression of active renin secretion is removed. In this setting, AII levels fall to the attomolar range (10−18moles), with AI and related metabolites being increased on the order of 100-fold, creating subsequent potential for cross-reactivity in radioimmunoassay for AII. Therefore, the different angiotensin peptides are separated before specific AI or II radioimmunoassay using cross-reacting antibodies labeled with 125I, with subsequent high-performance liquid chromatography using a octadecesilyl-silica stationary phase and changing (gradient) and constant (isocratic) mobile phases. The intraassay and interassay coefficients of variability were <10% for both assays.
Plasma ACE and aldosterone assays
Blood aliquots were taken into lithium heparin tubes, separated immediately and stored at −20°C until analysis. Plasma ACE was determined by the standard spectophotometric kinetic rate method, using the synthetic substrate N-3-(2-furyl) acryloyl-L-phenylalanylglycyl-glycine. Aldosterone assays were performed by an in-house radioimmunoassay using a standard commercial kit (Sorin Biomedica, Saluggia, Italy). The intraassay and interassay coefficients of variability were <9% for both assays.
Forearm blood flow (FBF) values were expressed as ml/min per 100 ml forearm volume. These blood flows were then converted to the ratio between the increase in blood flow in the infused arm and the blood flow in the control arm, according to the Whitney method (14).
Clinical characteristics between study visits were compared using Student’s paired ttests. Statistical analysis of FBF measurements for individual subjects were compared between study visits using two-way analysis of variance with repeated measures. The Bonferroni method for calculating 95% confidence intervals was the relevant multiple comparison range test used, and the interaction term for within-group effects in the dose-ranging study was the visit order, to look for observable carryover effect independent from the dose of treatment effect. A p value of <0.05 was considered significant and a value of <0.01 highly significant. The plethysmographic data were expressed as mean ± standard error of mean. In the dose-ranging study, there was no statistically significant independent relationship between visit order and FBF measurements (p = 0.79), in contrast to the marked influence of ACE inhibitor dose on FBF.
This statistical methodology has been validated as being most accurate in reflecting true differences in blood flow characteristics (10). The plethysmography technique itself is well suited to relatively small studies in adults, being able to detect a change of <20% with >90% power and p < 0.05 in studies of ∼20 individuals studied on separate occasions (15).
To see if there was any dissociation between plasma RAS activation within an individual and vascular tissue RAS activation in the same individual, we measured correlation coefficients between relevant parameters.
Table 1shows the basic demography of the patients in each of the three studies.
Forearm vascular blood flow responses to AIand AII in study I (cross-sectional study) (fig. 1)
A significant increase in AI-mediated vasoconstriction consistent with increased vascular AI/AII conversion and reduced vascular ACE inhibition was noted in those subjects with NYHA class III CHF as compared to those with NYHA class I or II CHF (maximum vasoconstriction −53 ± 6% [NYHAIII] versus −44 ± 5% [NYHA II] or −42 ± 5% [NYHA I]; p < 0.05 for NYHA class III vs. class I, difference between whole dose-response curves). In contrast, there were no differences observed for the control vasoconstrictor AII responses.
Forearm vascular blood flow responses to AI and AII in study II (longitudinal study) (fig. 2)
There was a consistent trend toward progressive activation of vascular AI/AII conversion with time with regards to AI-mediated vasoconstriction, with a significant increase in vasoconstriction observed between baseline and 18-month visits (maximum vasoconstriction elicited was −42 ± 3% [baseline] vs. −63 ± 2% [18-month visit], p< 0.001 for difference between dose-response curves). However, there was no significant difference among the three study visits regarding the forearm responses to the control vasoconstrictor AII (Fig. 2), nor were there any changes over time in their NYHA class or their doses of furosemide or lisinopril.
Forearm vascular responses to AI and AII with changes in ACE inhibitor dosing (fig. 3)
There was significant reduction in AI-mediated vasoconstriction on increasing the ACE inhibitor lisinopril dose from 10 mg/day to 20 mg/day, implying substantial reinhibition of vascular ACE (maximum vasoconstriction −64 ± 3% [10mg/day] versus −32 ± 5% [20 mg/day], p < 0.001 for difference for dose-response curves). There was no significant reduction in vascular ACE inhibition on halving the lisinopril dose to 5 mg/day (maximum vasoconstriction −69 ± 2%). There were no significant differences among the three study visits regarding the forearm responses to the control vasoconstrictor AII.
Differences in angiotensin peptides, aldosterone and ACE levels according to NYHA class (figs. 4–6)
⇑⇓⇓There were no significant differences in AI, aldosterone or plasma ACE levels between patients with different NYHA classes in the cross-sectional study (Fig. 4). However, there was a significant rise in AII levels observed between NYHA classes I/II and class III (6.1 ± 2.1 pg/ml [NYHA I], 6.3 ± 1.4 pg/ml [NYHA II] vs. 13.7 ± 2.0 pg/ml [NYHA III]; p = 0.005 for difference between NYHA classes II and III). Accordingly the AII/I ratio was also significantly elevated between NYHA classes I/II and III (0.04 ± 0.009 [NYHA II] vs. 0.07 ± 0.01 [NYHA III]; p = 0.03).
Changes in angiotensins, aldosterone and ACE levels longitudinally with time
There was a significant increase seen in AI levels between baseline and the two subsequent study visits, although the most dramatic rise was observed in the first nine months (169 ± 29 pg/ml [baseline], 255 ± 37 pg/ml [nine months], 278 ± 37 pg/ml; p = 0.02 between baseline and nine months, p = 0.006 between baseline and 18 months). There were, however, no significant changes in AII levels between study visits, nor were there any changes seen in AII/I ratios, aldosterone or plasma ACE levels (Fig. 5).
Angiotensin, aldosterone and ace with changing dose of ace inhibitor
There were no significant changes observed in any of the plasma parameters with either increasing or decreasing the dose of lisinopril for >6 weeks in the third study (Fig. 6).
Pressor dose responses to AI
The vascular tissue responses were also analyzed in a different way. In each of the three studies, each individual dose-response curve to AI was constructed to calculate for that individual the dose of AI required to reduce forearm blood flow by 30% (the AI-ΔFBF30%value), which would indicate a clinically significant change in blood flow related to vascular AI/AII conversion by ACE in vivo. These individual values were then averaged to look for differences among the three study groups (Fig. 7).
In the cross-sectional study, there was a significant reduction in AI-ΔFBF30%between NYHA I and III groups (p = 0.005), indicative of increased vascular ACE activity in the more severe disease class. We found a similar reduction seen between NYHA II and III (p = 0.013), although there was no significant difference between the milder NYHA I and II patient responses. In the longitudinal study, there was a significant reduction in AI-ΔFBF30%between baseline and 18 months (p = 0.01), implying increased vascular AI/AII conversion with progression of time and a nearly significant trend toward a similar increase in AI/AII conversion between baseline and nine months (p = 0.06). In the dose-ranging study, there was a dramatic attenuation in vascular AI/AII conversion between 5 mg and 20 mg/day (p < 0.0001), as well as between 10 mg and 20 mg/day (p = 0.0001), although not between 5 mg and 10 mg/day. The results thereafter are virtually identical whichever way the vascular tissue AI/AII conversion results are calculated (comparing Fig. 7with Figs. 1–3).
Correlation coefficients comparing vascular flow indices with plasma indices
There was no evidence of statistically significant correlation between the absolute plasma and vascular indices of RAS activity in any of the three studies performed (Table 2).
In the longitudinal study, correlations between the changes in AI-FBF30%values from baseline to 18 months against changes in plasma AII and aldosterone over the same time frame were calculated, but neither relationship was statistically significant (Pearson’s correlation coefficients: R2= 0.01 for aldosterone vs. AI-FBF30%, R2= 0.16 for AII vs. AI-FBF30%). In the dose-ranging study, similar correlation calculations were also not significant (R2= 0.13 for aldosterone vs. AI-FBF30%, R2= 0.08 for AII vs. AI-FBF30%). This is strong evidence of dissociation between plasma RAS activation and vascular tissue RAS activation.
Our main study findings fall into four categories. First, even in the presence of commonly used doses of ACE inhibitor therapy, vascular AI/AII conversion was greater in the severer cases of CHF. Second, even when the CHF disease process appears stable clinically, as in our longitudinal study, vascular AI/AII conversion increases gradually over time despite chronic ACE inhibitor therapy. These latter data are the most novel and the most important finding to arise from these studies. Third, increasing the dose of the ACE inhibitor markedly suppresses the reactivated vascular AI/AII conversion seen during chronic ACE inhibitor therapy. Fourth, interestingly, there was no correlation noted between any of the recognized markers of plasma RAS activity and the derived vascular ACE activity responses, which demonstrates for the first time in humans that there is dissociation between the plasma and tissue ACE systems in vivo. A larger sample size might have found statistical significance, but given the very low r2values we found, this seems unlikely.
We did consider measuring ACE genotypes to see if this determined AI/AII conversion, but there were two reasons against it. First, there were no obvious interindividual differences in vascular AI/AII conversion reactivation in our study, and second, a genotype study with three genotypes in only 28 individuals would be absurdly small and could lead to erroneous conclusions. This is an issue that needs to be studied more definitely in the future rather than be addressed by this study, which was really too small.
Experimental evidence suggests that blocking tissue ACE activity requires higher than normal doses of ACE inhibitors (16–18). In accordance with that finding, we found that plasma ACE did not reflect the reactivation we saw in vascular tissue AI/AII conversion. This was despite AI levels increasing between baseline and subsequent visits, which was presumably due to accumulation of AI “upstream” of the enzyme blockade.
Much has been written about which precise enzyme causes tissue AI/AII conversion. In humans in vivo, Schalekamp found that AI/AII conversion across organs was more than 90% ACE inhibitable (19). On the other hand, in ex vivo experiments, a role for chymase has been clearly demonstrated, particularly in the extravascular interstitium (9,20). Our studies did not attempt (nor were they designed) to directly address the question of which enzyme was most responsible, and therefore our results are not definitive in this regard. Our data are, however, consistent with previous data that ACE is solely responsible for AI/AII conversion in the intravascular space (21). However, our data do not exclude an additional contribution from chymase, as we did not perform studies with specific chymase inhibitor drugs. Our study was designed to answer the clinically relevant question of whether, as CHF progresses over time, vascular AI/AII conversion outstrips the ability of ACE inhibitors to prevent it. Clearly that is not the case. Therefore, even if chymase does contribute to some extent to overall vascular AI/AII conversion, ACE also contributes substantially, which means that increasing the ACE inhibitor dose would further suppress vascular AI/AII conversion in the clinical setting. Our study, however, does not assess whether adding an angiotensin receptor antagonist would achieve a better overall effect than increasing the dose of the ACE inhibitor. Our study merely suggests that a higher ACE inhibitor dose is one option, as suggested before (8).
The dose of lisinopril used here was the commonest dose used in the UK as well as being in the middle of the two doses used in the Assessment of Treatment with Lisinopril and Survival (ATLAS) trial and being above the dose of lisinopril used in the GISSI-3 study. Furthermore, the dose of lisinopril used here matches the average bioequivalent dose of enalapril used in the CONSENSUS I, SOLVD and RALES trials. We were keen to match perceived normal clinical practice, which explains our choice of drug and dosage studied.
In pharmacologic studies, clinically used ACE inhibitors vary in binding affinity for tissue-bound ACE (22–24). In terms of tissue binding, lisinopril is considered intermediate between the lipophilic highly tissue-bound ACE inhibitors such as quinapril and ramipril and the poorly tissue-bound ACE inhibitors such as captopril. However, because we were studying patients with CHF, we wanted to use the type of ACE inhibitors used in most of the major CHF trials. Clearly, it would be intriguing to repeat these studies with other ACE inhibitors that are more highly tissue bound, to see if vascular AI/AII conversion escapes as much with those agents as we found here with lisinopril.
An obvious question is the nature of the mechanism of the reactivated vascular AI/AII conversion we found during chronic ACE inhibitor therapy in CHF. Two main possibilities exist. First, Fyhrquist et al. (25)found in tissue culture that captopril induces ACE activity in human endothelial cells (25). This is akin to the phenomenon of upregulation of receptors when they are exposed constantly to an antagonist drug. It is quite possible that constant exposure to an ACE inhibitor causes induction of endothelial ACE activity as a homeostatic mechanism. The second possibility is that it is the progressive atherosclerotic process itself that is inducing ACE activity (26–28). Further work would be required to clarify these mechanisms more precisely.
What are the potential clinical consequences of these observations? The ATLAS trial found that high dose lisinopril was more beneficial overall than low dose lisinopril (29). High dose ACE inhibition produced a nonsignificant 8% (p = 0.128) reduction in mortality but a significant 24% reduction in hospitalizations for heart failure (p = 0.002). Intriguingly, recent data on the addition of an AII receptor antagonist to an ACE inhibitor has similarly produced a large effect on morbidity with little or no effect on mortality (30). For example, in the recent unpublished ValHeFT trial, valsartan dramatically reduced hospitalizations without altering total mortality. A picture is therefore developing that AII reactivation during chronic ACE inhibitor therapy mainly has an adverse effect on morbidity rather than on mortality. The results we found here raise the possibility that ACE inhibitor doses should be gradually increased during chronic therapy.
In conclusion, vascular AI/AII conversion reactivates despite conventional doses of chronic lisinopril therapy in patients with CHF. This occurs over time even if the CHF disease process is clinically apparently stable, but it also occurs as the disease process progresses. However, this reactivation of AI/AII conversion is not purely due to non-ACE pathways. The reactivation is still ACE inhibitor suppressible because increasing the dose of the ACE inhibitor suppresses vascular tissue AI/AII conversion markedly even after the conversion has reactivated. Finally, there would appear to be a total dissociation between reactivation of the plasma RAS and vascular AI/AII conversion.
☆ C. A. J. Farquharson was supported by the British Heart Foundation during this study.
- angiotensin-converting enzyme
- angiotensin I
- angiotensin II
- Assessment of Treatment with Lisinopril and Survival
- chronic heart failure
- forearm blood flow
- New York Heart Association
- renin angiotensin system
- Received September 10, 2001.
- Revision received November 7, 2001.
- Accepted December 17, 2001.
- American College of Cardiology Foundation
- MacFadyen R.J,
- Lee A.F.C,
- Morton J.J,
- Pringle S.D,
- Struthers A.D
- Swedberg K,
- et al.
- Hirsch A.T,
- Talsness C.E,
- Schunkert H,
- et al.
- Massie B.M,
- Kramer B.L,
- Topic N
- Jorde U.P,
- Ennezat P.V,
- Lisker J,
- et al.
- Petrie M.C,
- Padmanabhan N,
- McDonald J.E,
- et al.
- Farquharson C.A.J,
- Struthers A.D
- Davidson N.C,
- Barr C.S,
- Struthers A.D
- Unger T,
- Ganten D,
- Lang R.E,
- Scholkens B.A
- Rakugi H,
- Wang D.S,
- Dzau V.J,
- Pratt R.E
- Wolny A,
- Clozel J.-P,
- Rein J,
- et al.
- Johnston C.I,
- Fabris B,
- Yamada H,
- et al.
- Fabris B,
- Chen B,
- Pupic V,
- Perich R,
- Johnston C.I
- Hoshida S,
- Kato J,
- Nishino M,
- et al.
- Packer M,
- Poole-Wilson P.A,
- Armstrong P.W,
- et al.
- Hamroff G,
- Katz S.D,
- Mancini D,
- et al.