1. 95th Annual Meeting: American Association for Thoracic Surgery
Late survival and RV performance in matched children after Norwood: Norwood-BT vs Norwood-Sano
Travis Wilder Brian McCrindle Alistair Phillips Eugene Blackstone Rajeswaran Jeevanantham William Williams William DeCampli
Jeffery Jacobs
Marshall Jacobs
Tara Karamlou
Paul Kirshbom
Gary Lofland
Gerhard Ziemer
Edward Hickey
Good morning, Dr. Ohye, members and guests
95th Annual Meeting: American Association for Thoracic Surgery
2015, Seattle, WA

2. No disclosures

3. Congenital Heart Surgeons’ Society
N = 692 Neonates
Critical Left Ventricular Outflow Tract Obstruction
2005 – 2014: 21 Institutions
Aortic atresia/critical AS
Ductal-dependent
The Congenital heart surgeons society has enrolled 692 neonates with critical Left ventricular outflow tract obstruction into a research database since 2005…

4. Congenital Heart Surgeons’ Society
N = 692 Neonates
Critical Left Ventricular Outflow Tract Obstruction
2005 – 2014: 21 Institutions
Aortic atresia/critical AS
Ductal-dependent
Aortic Valvotomy
n = 107
Ross-Konno/
Yasui
n = 8
Transplant
n = 5
Norwood
n = 454
Hybrid
n = 110
and because this is an all-inclusive inception cohort, covering a broad spectrum of disease, neonates were managed with a variety of initial procedures; however, the majority – 454 – underwent a stage-1 Norwood operation…

5. Stage-1 Norwood n = 454 BT (51%) n = 232 SANO (49%) n = 222
Ohye RG, et al. NEJM, 362;21 May, 2010
Ohye RG, et al. NEJM, 362;21 May, 2010
…which included 232 Norwood-BT shunt operations and 222 Norwood operations with Sano modifications … and recently the Sano operation has been shown to have improved early survival
However, the enthusiasm for the Sano operation has somewhat diminished because of concerns surrounding long-term ventricular dysfunction

6. BT vs SANO
Difference in survival Difference in RV dysfunction Difference in tricuspid regurgitation
So we set out to determine the difference is long-term survival, RV dysfunction and TR

7. Overall survival for all children
SANO
71%
59%
% Survival
P < 0.001 BT
Among all children, the overall, unadjusted survival was superior after a Sano; however, comparing survival among all children from each group doesn’t adjusted for potential baseline differences between the groups… so in order to try and compare apples to apples our analysis focused on a subset of propensity-matched children…
Years after Norwood

8. Propensity-Matching 100%
BT vs SANO
100%
And to briefly conceptualize this strategy, consider the red boxes represent all children who had a BT, blue boxes are children who had a SANO and the shapes are variations in baseline characteristics … Using a propensity-score, generated from these baseline characteristics we identified, as close as possible similar children in each group...

9. Matched Children BT (75%) n = 169 SANO (74%) n = 169 n = 338
which subsequently resulted in 169 matched-pairs, or about three-quarters of all children who underwent a Norwood…

10. Age at operation (days) 6.5 ± 5 6.3 ± 5 .14 Birth weight (kg)
BT (169)
SANO (169)
Variable
Value P
Age at operation (days)
6.5 ± 5
6.3 ± 5
.14
Birth weight (kg)
3.2 ± 0.4
3.2 ± 0.5
.92
Birth weight < 2.5 kg
9 (5%)
11 (7%)
.61
BSA at operation (m2)
0.21 ± 0.02
.71
Aortic valve annulus (cm)
0.15 ± 0.2
0.14 ± 0.2
.82
Aortic valve atresia
90 (55%)
89 (57%)
.74
TV annulus (cm)
1.3 ± .2
.77
MV annulus (cm)
.24 ± .3
.22 ± .4
.44
Mitral valve atresia
70 (45%)
75 (48%) AI
22 (15%)
19 (13%)
Transverse arch coarctation
52 (32%)
50 (30%)
.81
Small/restrictive ASD
54 (44%)
49 (41%)
.67
Important RV dysfunction
2 (1%)
.97
Important TR
10 (6%)
14 (9%)
.42
Renal insufficiency
1 (1%)
2 ( 1%)
.56
Metabolic acidosis
20 (12%) 1
…and this list of baseline characteristics…

11. Age at operation (days) 6.5 ± 5 6.3 ± 5 .14 Birth weight (kg)
BT (169)
SANO (169)
Variable
Value P
Age at operation (days)
6.5 ± 5
6.3 ± 5
.14
Birth weight (kg)
3.2 ± 0.4
3.2 ± 0.5
.92
Birth weight < 2.5 kg
9 (5%)
11 (7%)
.61
BSA at operation (m2)
0.21 ± 0.02
.71
Aortic valve annulus (cm)
0.15 ± 0.2
0.14 ± 0.2
.82
Aortic valve atresia
90 (55%)
89 (57%)
.74
TV annulus (cm)
1.3 ± .2
.77
MV annulus (cm)
.24 ± .3
.22 ± .4
.44
Mitral valve atresia
70 (45%)
75 (48%) AI
22 (15%)
19 (13%)
Transverse arch coarctation
52 (32%)
50 (30%)
.81
Small/restrictive ASD
54 (44%)
49 (41%)
.67
Important RV dysfunction
2 (1%)
.97
Important TR
10 (6%)
14 (9%)
.42
Renal insufficiency
1 (1%)
2 ( 1%)
.56
Metabolic acidosis
20 (12%) 1
…shows a similar baseline profile between groups…

12. Age at operation (days) 6.5 ± 5 6.3 ± 5 .14 Birth weight (kg)
BT (169)
SANO (169)
Variable
Value P
Age at operation (days)
6.5 ± 5
6.3 ± 5
.14
Birth weight (kg)
3.2 ± 0.4
3.2 ± 0.5
.92
Birth weight < 2.5 kg
9 (5%)
11 (7%)
.61
BSA at operation (m2)
0.21 ± 0.02
.71
Aortic valve annulus (cm)
0.15 ± 0.2
0.14 ± 0.2
.82
Aortic valve atresia
90 (55%)
89 (57%)
.74
TV annulus (cm)
1.3 ± .2
.77
MV annulus (cm)
.24 ± .3
.22 ± .4
.44
Mitral valve atresia
70 (45%)
75 (48%) AI
22 (15%)
19 (13%)
Transverse arch coarctation
52 (32%)
50 (30%)
.81
Small/restrictive ASD
54 (44%)
49 (41%)
.67
Important RV dysfunction
2 (1%)
.97
Important TR
10 (6%)
14 (9%)
.42
Renal insufficiency
1 (1%)
2 ( 1%)
.56
Metabolic acidosis
20 (12%) 1
Including key variables such as birth weight

13. Age at operation (days) 6.5 ± 5 6.3 ± 5 .14 Birth weight (kg)
BT (169)
SANO (169)
Variable
Value P
Age at operation (days)
6.5 ± 5
6.3 ± 5
.14
Birth weight (kg)
3.2 ± 0.4
3.2 ± 0.5
.92
Birth weight < 2.5 kg
9 (5%)
11 (7%)
.61
BSA at operation (m2)
0.21 ± 0.02
.71
Aortic valve annulus (cm)
0.15 ± 0.2
0.14 ± 0.2
.82
Aortic valve atresia
90 (55%)
89 (57%)
.74
TV annulus (cm)
1.3 ± .2
.77
MV annulus (cm)
.24 ± .3
.22 ± .4
.44
Mitral valve atresia
70 (45%)
75 (48%) AI
22 (15%)
19 (13%)
Transverse arch coarctation
52 (32%)
50 (30%)
.81
Small/restrictive ASD
54 (44%)
49 (41%)
.67
Important RV dysfunction
2 (1%)
.97
Important TR
10 (6%)
14 (9%)
.42
Renal insufficiency
1 (1%)
2 ( 1%)
.56
Metabolic acidosis
20 (12%) 1
Aortic valve atresia

14. Age at operation (days) 6.5 ± 5 6.3 ± 5 .14 Birth weight (kg)
BT (169)
SANO (169)
Variable
Value P
Age at operation (days)
6.5 ± 5
6.3 ± 5
.14
Birth weight (kg)
3.2 ± 0.4
3.2 ± 0.5
.92
Birth weight < 2.5 kg
9 (5%)
11 (7%)
.61
BSA at operation (m2)
0.21 ± 0.02
.71
Aortic valve annulus (cm)
0.15 ± 0.2
0.14 ± 0.2
.82
Aortic valve atresia
90 (55%)
89 (57%)
.74
TV annulus (cm)
1.3 ± .2
.77
MV annulus (cm)
.24 ± .3
.22 ± .4
.44
Mitral valve atresia
70 (45%)
75 (48%) AI
22 (15%)
19 (13%)
Transverse arch coarctation
52 (32%)
50 (30%)
.81
Small/restrictive ASD
54 (44%)
49 (41%)
.67
Important RV dysfunction
2 (1%)
.97
Important TR
10 (6%)
14 (9%)
.42
Renal insufficiency
1 (1%)
2 ( 1%)
.56
Metabolic acidosis
20 (12%) 1
And baseline RV dysfunction and TR… So we are in fact comparing apples to apples.

15. 1. Parametric hazard analysis: a. Competing end-states b
1. Parametric hazard analysis: a. Competing end-states b. Overall survival c. Transplant-free survival
We used parametric hazard analysis to evaluate transition to various competing end-states, overall survival and transplant-free survival

16. End-states after Norwood All Matched Children (n=338)%
Initially end-states were evaluated for 338 propensity matched children
Years after Norwood

17. End-states after Norwood All Matched Children (n=338)
Alive without
definitive palliation %
Immediately after Norwood all children are alive without further intervention – represented by black curve at time-zero, and thereafter transition into various mutually exclusive end-states. 9%
Years after Norwood

18. End-states after Norwood All Matched Children (n=338)
Alive without
definitive palliation %
By 6-years a small percentage of children underwent biventricular repair transplantation
Transplant
2-V 9% 4% 2%
Years after Norwood

19. End-states after Norwood All Matched Children (n=338)
Alive without
definitive palliation
Fontan
52% %
Just over half reached a Fontan
Transplant
2-V 9% 4% 2%
Years after Norwood

20. End-states after Norwood All Matched Children (n=338)
Alive without
definitive palliation
Fontan
52% %
Dead
33%
And a third of children died
Transplant
2-V 9% 4% 2%
Years after Norwood

21. End-states after Norwood
Mortality
BT (n=169)
Death after BT %
42%
When end-states were evaluated independently for each operation type, children who underwent BT had a 42% mortality before a definitive palliation
Years after Norwood

22. End-states after Norwood
Mortality
SANO (n=169)
Death after BT %
42%
Death after SANO
Compared to children who underwent a SANO who only had a 24% mortality…
24%
Years after Norwood

23. End-states after BT (n=169)
Definitive Palliation
Fontan %
49%
Furthermore, children who underwent a BT were less likely to transition to definitive palliation
Transplant/2-V 3%
Years after Norwood

24. End-states after SANO (n=169)
Definitive Palliation
Fontan
Fontan
54% %
49%
Transplant/2-V
Compared to children who underwent a SANO… and although this wasn’t statistically different, it does imply that these children are more likely to reach a definitive palliation - irrespective of the final strategy.
Transplant/2-V
10% 3%
Years after Norwood

25. Overall survival n = 338 SANO BT % Survival Years after Norwood 70%
55%
% Survival
P < 0.001 BT
And that is reflected in this overall, survival cure, now for only propensity matched children, with this survival curve showing that the overall survival was significantly better after Sano – regardless of where they end up
Years after Norwood

26. Transplant-free survival
SANO
64%
53%
% Survival
P = 0.004 BT
Similarly, when we compared transplant-free survival – although there was a late convergence between the survival curves, children who underwent a Sano had a better freedom from death or transplant at all points in time during the 6-year follow-up.
Years after Norwood

27. SANO Systemic RV Ventriculotomy
Ohye RG, et al. NEJM, 362;21 May, 2010
Systemic RV
A primary objection to the Sano, is that an incision (or ventriculotomy) is required to the systemic ventricle, which theoretically compromises late ventricular function – and perhaps tricuspid valve performance
Ventriculotomy

28. 2. Prevalence of RV dysfunction and TR: Mixed-model regression analysis of repeated echos
Therefore we examined the prevalence of RV dysfunction to and tricuspid valve regurgitation between the 2 surgical options…
The next portion of our analysis focused on the prevalence of RV dysfunction and tricuspid regurgitation.

29. Reports analyzed from every echo on every patient
2. Prevalence of RV dysfunction and TR: Mixed-model regression analysis of repeated echos
And did so by using mixed-model regression to analyzing reports from every echo we could acquire for all children, throughout follow-up.
Reports analyzed from every echo on every patient
N = 2,993 echos

30. Grades of RV dysfunction and TR From institutional echo reports
1=trivial
2=mild
3=mild-to-moderate
4=moderate
5=moderate-to-severe
6=severe
Initially, RV dysfunction and TR were graded from trivial to severe using a 6 point scale, according to the reports…

31. Grades of RV dysfunction and TR From institutional echo reports
1=trivial
2=mild
3=mild-to-moderate
4=moderate
5=moderate-to-severe
6=severe
“Important”
RV dysfunction
and TR
However, we considered “moderate or worse” to be clinically important dysfunction and regurgitation, and this is the threshold we used for our analysis.

32. RV dysfunction after Norwood
2474 echos for 292 children
≥ Moderate RV dysfunction (%)
We first determined prevalence of important RV dysfunction for all propensity matched children after Norwood… which is represented by the blue curve which demonstrating the change in prevalence of important RV dysfunction on the vertical axis over time after Norwood operation – on the horizontal axis; and for all survivors, important RV dysfunction peaks to nearly 12% within the initial 6 months, after which there is a decline to low, stable, late levels…
Years after Norwood

33. RV dysfunction after Norwood
BT vs SANO
Early ,P = 0.019 BT
≥ Moderate RV dysfunction (%)
SANO
We then stratified the prevalence of RV dysfunction by the type of Norwood, and children who underwent BT had a significant increase in the early prevalence of RV dysfunction, nearly 3 times that of the Sano operation despite similar preoperative RV dysfunction between both groups - Thereafter, however, both groups experienced a declining prevalence, resulting in a non-significant difference in late RV dysfunction.
Late, P = 0.36
Years after Norwood

34. Transplant-free survival and RV dysfunction
BT vs SANO
Transplant-free survival: BT vs SANO
64%
53% %
To get a better idea of the relationship between survival and RV dysfunction we plotted transplant-free survival over the same time…
Years after Norwood

35. Transplant-free survival and RV dysfunction
BT vs SANO
Transplant-free survival: BT vs SANO
64%
53% %
RV dysfunction:
BT vs SANO
…as the prevalence of RV dysfunction
Years after Norwood

36. Transplant-free survival and RV dysfunction
BT vs SANO
Transplant-free survival: BT vs SANO
64%
53% %
Early hazard
RV dysfunction:
BT vs SANO
And not surprisingly, the early hazard for death, coincided temporally with the peak prevalence of RV dysfunction… and although this is rather intuitive, is suggestive of a causal link between RV dysfunction and survival… and furthermore it suggests the morbidity and mortality of a Norwood occur early…
Peak prevalence
Years after Norwood

37. Transplant-free survival after stage-2
BT (n=108) vs SANO (n=125)
81%
80%
SANO BT
P = 0.68
% Survival
Which is emphasized by comparing transplant-free after transition to a stage-2 cavopulmonary shunt operation… and survival conditional on achieving stage-2 – which is now time-zero on this figure – was not different between BT or SANO…
Years after Stage-2

38. RV dysfunction after stage-2
BT vs SANO
SANO
≥ Moderate RV dysfunction (%)
Similarly, there wasn’t a difference in RV dysfunction after stage-2, and this emphasizes that the cost of Norwood - and specifically a BT - occurs early; However, after transition to a staged ventricular unloading operation, survival and RV dysfunction are stable for both groups.
P = 0.675 BT
Years after Stage-2

39. Tricuspid regurgitation after Norwood
2423 echos for 288 children
≥ Moderate TR (%)
Finally, switching our attention to trends in tricuspid regurgitation…
Notice that clinically important TR generally affects a higher proportion of children - about 10-15% - throughout follow-up, and that the time-related variations are less pronounced than that of RV dysfunction…
Years after Norwood

40. Tricuspid regurgitation after Norwood
BT vs SANO
Early ,P = 0.003 BT
≥ Moderate TR (%)
Late, P = 0.68
SANO
And when TR was stratified by type of Norwood, BT operations were associated with a significant early increase in TR, however late TR was similar between groups.
Years after Norwood

41. Summary
For comparable neonates with critical LVOTO undergoing stage-1 Norwood:
Overall and transplant-free survival are better after SANO

42. Summary
For comparable neonates with critical LVOTO undergoing stage-1 Norwood:
Overall and transplant-free survival are better after SANO
Early RV dysfunction and TR are worse after BT

43. Summary
For comparable neonates with critical LVOTO undergoing stage-1 Norwood:
Overall and transplant-free survival are better after SANO
Early RV dysfunction and TR are worse after BT
Late RV dysfunction and TR are similar between groups

44. Implication
Children undergoing SANO are more likely to be alive at 6 years – irrespective of final physiology Suggesting that SANO is preferable to BT for stage-1 palliation