Marfan Sartan: A Randomized Placebo-Controlled Trial
Marfan Sartan: A Randomized Placebo-Controlled Trial
The Marfan Sartan trial was designed to evaluate the benefit of adding Losartan to baseline therapy in patients with MFS. This randomized, double-blind, multi-centre trial demonstrates that Losartan does not significantly alter the aortic root dilatation rate in this population. This result was observed despite the slight but significant decrease in both systolic and diastolic blood pressure in patients receiving Losartan, indicating that the expected haemodynamic effect of Losartan was observed.
Actually, the benefit of Losartan reported in studies of patients with Marfan appears to decrease as time progresses, and as the size of the populations studied increases: the first retrospective study of 18 children, mean age 6.5 years, selected at the time of maximal dilatation without a control group reported a 86% decrease in aortic dilatation rate; similarly, a randomized unblinded study comparing 28 children (mean age 13.1 years) receiving Losartan on top of β-blocker therapy reported a 0.79 mm/year lower aortic dilatation rate in the group receiving Losartan (0.1 mm/year vs. 0.89 mm/year), but baseline aortic diameter differed between the two groups.
Groenink reported a smaller 0.19 mm/year mean difference in aortic dilatation rate in favour of Losartan. This difference was measured in the subgroup of 137 patients (76 Losartan, 61 controls) with native aorta at the end of the study, i.e. 62% of randomized patients (whose mean age was 37.5 years) which may have led to a bias. Actually, the baseline aortic diameter was lower in the control group that in the group of patients receiving Losartan. Of note, the dilatation rate was higher in the control group than that reported previously by the same group with echocardiography in Marfan patients (0.63 vs. 0.40 mm/year). A smaller aortic dilatation rate was measured with MRI for unclear reasons. The number of prophylactic aortic surgery was similar in the two groups (10 Losartan, 8 control group).
Finally, very recently, the Pediatric Heart Network Study, designed as a superiority trial, failed to demonstrate the superiority of Losartan over Atenolol in limiting aortic dilatation in 608 children and young adults (mean age 11.5 years) with MFS. Actually, in the group of patients receiving Losartan, the number of aortic events tended to be greater (19 vs. 10), and the aortic dilatation also tended to be greater, a difference which reached statistical significance at the level of the aortic annulus.
All studies evaluating Losartan used the Ghent 1 criteria to select the population. The use of the new nosology proposed in 2010 would most likely not have modified the results: comparison of the two criteria in probands carrying an FBN1 gene mutation suggested that aortic dilatation was slightly more prevalent with the use of the recent criteria, and a high level of agreement between the two sets of criteria was reported in a Korean population with or without an FBN1 gene mutation present.
It is unlikely that our trial misleadingly concludes to the absence of any clinically relevant effect of Losartan in patients with MFS for several reasons : (i) our study is the largest reported to date directly evaluating the effect of Losartan in Marfan patients; (ii) the expected number of patients was included, although only 146 and 147 patients could be evaluated for aortic dilatation in each group; (iii) the duration of follow-up was longer than in previous studies (median 3.5 years); (iv) the severity of the population was similar to that included in other studies as indicated by a baseline z-zcore (3.7) similar to that in Groenink's study (3.8) and the number of clinical events (9.5%) similar to that reported in other studies (8.5% in the Atenolol vs. Losartan trial and 8.7% in Groenink's study); (v) our results are very consistent whatever the section of the aorta being considered, and whatever the subgroup considered as there is no interaction with age, baseline aortic diameter, presence or absence of an FBN1 mutation, presence or absence of β-blockade; (vi) our statistical analysis reinforces the chance of getting a positive study beyond classical per-protocol evaluation: the use of the slope of aortic diameter evolution derived from multiple echocardiographic measurements allows both improved precision in the aortic dilatation rate because of the multiple measures, and inclusion of data under treatment of the patients who later discontinued the drug or the placebo for whatever reason (including surgery), while the patient was still receiving the drug or the placebo.
The absence of significant benefit from Losartan in patients with MFS may indicate that the dosage of Losartan was too low in our study, as in the mouse model, which suggested the benefit of Losartan, the dosage given was tapered so as to decrease blood pressure and heart rate by 15–20%, and was probably much higher. However, 100 mg was also the maximal dose chosen for the other studies because it is the maximal recommended dose of the drug in humans for hypertension. Furthermore, a 5 mmHg significant decrease in blood pressure was observed in our normotensive population with few instances of symptomatic hypotension thus suggesting that the effective dosage was obtained. It is also conceivable that a beneficial effect is only possible in the very early stage, as in the mouse model. Losartan was given first during pregnancy and the experiment repeated with Losartan given at 7 weeks after echocardiographic documentation of aneurysm. In our study, no indication of benefit of Losartan was observed in the subgroup of patients without dilated aorta.
The negative results of our trial may be attributed to the importance of complete blinding in randomized trials as recently outlined by the renal denervation studies in hypertension. Indeed, initial studies, incompletely blinded, reported a significant decrease in blood pressure associated with the procedure, whereas the last and completely blinded study demonstrated the absence of effect of denervation on blood pressure in hypertensive patients. The opposite results in the Losartan trials in MFS similarly illustrate the importance of full blinding before a definitive conclusion can be reached. The progressive decrease in the benefit observed with Losartan in previous trials also supports this hypothesis (from 3.08 mm/year in Brooke study with historical control, 0.79 mm/year in Chiu's study and 0.19 mm/year in Groenink's study both with a control group) and questioned the clinical significance of this effect.
The absence of efficacy of Losartan also questions the applicability in humans with MFS of the TGFβ hypothesis as proposed in the mouse model. The rationale for evaluating the effect of Losartan in MFS patients was based on the Fbn1 KI mouse model (Fbn1), in which aortic dilatation was stopped or partly prevented by Losartan treatment. Our negative results, the limited benefit reported by Groenink, the absence of superiority over Atenolol in the Pediatric Heart Network Study, would all suggest that the applicability of mouse data in human remains unclear.
The vasodilator properties of Losartan could have been expected to be beneficial in Marfan patients by producing a lowering in mean blood pressure and a decrease and delay of the rebound wave. Both contribute to lower central systolic blood pressure which has been associated with aortic dilatation in this population. Angiotensin receptor blocker appear to be more efficacious in lowering central systolic pressure than β-blockers in hypertensive patients, and the effects of β-blockers on aortic compliance in patients with MFS appear to be variable. However, the alteration of vessel walls in MFS patients (indicated by decreased aortic wall compliance and increased pulse wave velocity) could have blunted the vasodilator action of Losartan. It has been suggested that the nitric oxide pathway was not impacted by Losartan and that may explain long-term failure. Besides, it is known that the vasodilator properties of Losartan are limited in normotensive patients such as those included in our study. In fact, a slight but significant decrease in peripheral blood pressure was observed in the Losartan group (Figure 4), suggesting that the expected haemodynamic effect was obtained.
β-Blocker therapy has been demonstrated to be efficacious in limiting aortic root dilatation in patients with MFS in an open label randomized study including 70 patients published 20 years ago. Since then, only retrospective studies including few patients have been performed producing contrasting results, some positive, some negative. It remains nevertheless the rule to propose β-blocker therapy in patients with MFS.
In conclusion, the evolution of aortic diameter at the level of the sinuses of Valsalva in Marfan patients was not modified by the adjunction of Losartan to baseline therapy.
The practical consequence of our study is that Losartan should not be systematically proposed with MFS, as there is no evidence that Losartan may be of benefit in this population. β-Blocker therapy (which was taken by over 80% of our patients) remains the first-line therapy that should be proposed to these patients, and possibly to patients with AAAs from other aetiologies.
Discussion
The Marfan Sartan trial was designed to evaluate the benefit of adding Losartan to baseline therapy in patients with MFS. This randomized, double-blind, multi-centre trial demonstrates that Losartan does not significantly alter the aortic root dilatation rate in this population. This result was observed despite the slight but significant decrease in both systolic and diastolic blood pressure in patients receiving Losartan, indicating that the expected haemodynamic effect of Losartan was observed.
Actually, the benefit of Losartan reported in studies of patients with Marfan appears to decrease as time progresses, and as the size of the populations studied increases: the first retrospective study of 18 children, mean age 6.5 years, selected at the time of maximal dilatation without a control group reported a 86% decrease in aortic dilatation rate; similarly, a randomized unblinded study comparing 28 children (mean age 13.1 years) receiving Losartan on top of β-blocker therapy reported a 0.79 mm/year lower aortic dilatation rate in the group receiving Losartan (0.1 mm/year vs. 0.89 mm/year), but baseline aortic diameter differed between the two groups.
Groenink reported a smaller 0.19 mm/year mean difference in aortic dilatation rate in favour of Losartan. This difference was measured in the subgroup of 137 patients (76 Losartan, 61 controls) with native aorta at the end of the study, i.e. 62% of randomized patients (whose mean age was 37.5 years) which may have led to a bias. Actually, the baseline aortic diameter was lower in the control group that in the group of patients receiving Losartan. Of note, the dilatation rate was higher in the control group than that reported previously by the same group with echocardiography in Marfan patients (0.63 vs. 0.40 mm/year). A smaller aortic dilatation rate was measured with MRI for unclear reasons. The number of prophylactic aortic surgery was similar in the two groups (10 Losartan, 8 control group).
Finally, very recently, the Pediatric Heart Network Study, designed as a superiority trial, failed to demonstrate the superiority of Losartan over Atenolol in limiting aortic dilatation in 608 children and young adults (mean age 11.5 years) with MFS. Actually, in the group of patients receiving Losartan, the number of aortic events tended to be greater (19 vs. 10), and the aortic dilatation also tended to be greater, a difference which reached statistical significance at the level of the aortic annulus.
All studies evaluating Losartan used the Ghent 1 criteria to select the population. The use of the new nosology proposed in 2010 would most likely not have modified the results: comparison of the two criteria in probands carrying an FBN1 gene mutation suggested that aortic dilatation was slightly more prevalent with the use of the recent criteria, and a high level of agreement between the two sets of criteria was reported in a Korean population with or without an FBN1 gene mutation present.
It is unlikely that our trial misleadingly concludes to the absence of any clinically relevant effect of Losartan in patients with MFS for several reasons : (i) our study is the largest reported to date directly evaluating the effect of Losartan in Marfan patients; (ii) the expected number of patients was included, although only 146 and 147 patients could be evaluated for aortic dilatation in each group; (iii) the duration of follow-up was longer than in previous studies (median 3.5 years); (iv) the severity of the population was similar to that included in other studies as indicated by a baseline z-zcore (3.7) similar to that in Groenink's study (3.8) and the number of clinical events (9.5%) similar to that reported in other studies (8.5% in the Atenolol vs. Losartan trial and 8.7% in Groenink's study); (v) our results are very consistent whatever the section of the aorta being considered, and whatever the subgroup considered as there is no interaction with age, baseline aortic diameter, presence or absence of an FBN1 mutation, presence or absence of β-blockade; (vi) our statistical analysis reinforces the chance of getting a positive study beyond classical per-protocol evaluation: the use of the slope of aortic diameter evolution derived from multiple echocardiographic measurements allows both improved precision in the aortic dilatation rate because of the multiple measures, and inclusion of data under treatment of the patients who later discontinued the drug or the placebo for whatever reason (including surgery), while the patient was still receiving the drug or the placebo.
The absence of significant benefit from Losartan in patients with MFS may indicate that the dosage of Losartan was too low in our study, as in the mouse model, which suggested the benefit of Losartan, the dosage given was tapered so as to decrease blood pressure and heart rate by 15–20%, and was probably much higher. However, 100 mg was also the maximal dose chosen for the other studies because it is the maximal recommended dose of the drug in humans for hypertension. Furthermore, a 5 mmHg significant decrease in blood pressure was observed in our normotensive population with few instances of symptomatic hypotension thus suggesting that the effective dosage was obtained. It is also conceivable that a beneficial effect is only possible in the very early stage, as in the mouse model. Losartan was given first during pregnancy and the experiment repeated with Losartan given at 7 weeks after echocardiographic documentation of aneurysm. In our study, no indication of benefit of Losartan was observed in the subgroup of patients without dilated aorta.
The negative results of our trial may be attributed to the importance of complete blinding in randomized trials as recently outlined by the renal denervation studies in hypertension. Indeed, initial studies, incompletely blinded, reported a significant decrease in blood pressure associated with the procedure, whereas the last and completely blinded study demonstrated the absence of effect of denervation on blood pressure in hypertensive patients. The opposite results in the Losartan trials in MFS similarly illustrate the importance of full blinding before a definitive conclusion can be reached. The progressive decrease in the benefit observed with Losartan in previous trials also supports this hypothesis (from 3.08 mm/year in Brooke study with historical control, 0.79 mm/year in Chiu's study and 0.19 mm/year in Groenink's study both with a control group) and questioned the clinical significance of this effect.
The absence of efficacy of Losartan also questions the applicability in humans with MFS of the TGFβ hypothesis as proposed in the mouse model. The rationale for evaluating the effect of Losartan in MFS patients was based on the Fbn1 KI mouse model (Fbn1), in which aortic dilatation was stopped or partly prevented by Losartan treatment. Our negative results, the limited benefit reported by Groenink, the absence of superiority over Atenolol in the Pediatric Heart Network Study, would all suggest that the applicability of mouse data in human remains unclear.
The vasodilator properties of Losartan could have been expected to be beneficial in Marfan patients by producing a lowering in mean blood pressure and a decrease and delay of the rebound wave. Both contribute to lower central systolic blood pressure which has been associated with aortic dilatation in this population. Angiotensin receptor blocker appear to be more efficacious in lowering central systolic pressure than β-blockers in hypertensive patients, and the effects of β-blockers on aortic compliance in patients with MFS appear to be variable. However, the alteration of vessel walls in MFS patients (indicated by decreased aortic wall compliance and increased pulse wave velocity) could have blunted the vasodilator action of Losartan. It has been suggested that the nitric oxide pathway was not impacted by Losartan and that may explain long-term failure. Besides, it is known that the vasodilator properties of Losartan are limited in normotensive patients such as those included in our study. In fact, a slight but significant decrease in peripheral blood pressure was observed in the Losartan group (Figure 4), suggesting that the expected haemodynamic effect was obtained.
β-Blocker therapy has been demonstrated to be efficacious in limiting aortic root dilatation in patients with MFS in an open label randomized study including 70 patients published 20 years ago. Since then, only retrospective studies including few patients have been performed producing contrasting results, some positive, some negative. It remains nevertheless the rule to propose β-blocker therapy in patients with MFS.
In conclusion, the evolution of aortic diameter at the level of the sinuses of Valsalva in Marfan patients was not modified by the adjunction of Losartan to baseline therapy.
The practical consequence of our study is that Losartan should not be systematically proposed with MFS, as there is no evidence that Losartan may be of benefit in this population. β-Blocker therapy (which was taken by over 80% of our patients) remains the first-line therapy that should be proposed to these patients, and possibly to patients with AAAs from other aetiologies.
