Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Alteration of the critical arteriovenous oxygen saturation relationship by sustained afterload reduction after the Norwood procedure.

Hoffman GM, Tweddell JS, Ghanayem NS, Mussatto KA, Stuth EA, Jaquis RD, Berger S.

J Thorac Cardiovasc Surg. 2004 Mar;127(3):738-45. [abstract]

Reviewed by Nancy Lin MD, Pediatric Cardiology and Critical Care, St. Louis Children's Hospital, Washington University School of Medicine

Review posted May 6, 2005

---

I. What is being studied?:

The study objective:

1. To confirm the existence of a critical range of arterial oxygen saturations (SaO2) for optimizing systemic oxygen delivery/maximizing mixed venous saturation (SvO2) in patients undergoing stage I palliation (Norwood procedure) for hypoplastic left heart syndrome (HLHS).
2. To test the hypothesis that this SaO2-SvO2 relationship could be altered by afterload reduction with phenoxybenzamine.

The study design:

Analysis of data collected prospectively from a nonrandomized "surgeon-directed protocol"

The patients included:

All patients undergoing the Norwood procedure at the Children's Hospital of Wisconsin from July 1996 to July 2000 were included.

The patients excluded:

None.

The interventions compared:

On initiation of cardiopulmonary bypass (CPB), phenoxybenzamine 0.25 mg/kg was given to 62 of 71 neonates undergoing the Norwood procedure. Among infants who received phenoxybenzamine on CPB, those who demonstrated low SvO2 with elevated Qp:Qs in the postoperative period were administered phenoxybenzamine (10 mcg/kg/hr) as adjuvant therapy to lower SVR. The other 9 patients who did not receive the drug served as controls.

The outcomes evaluated:

Primary outcomes: arterial saturations by pulse oximetry, oximetric SvO2 measurements from SVC catheters, and the resulting arteriovenous oxygen content differences (ΔCa-vO2) (using measured hemoglobins and an assumed oxygen consumption of 160 ml/min/m2) were monitored hourly for the first 48 post-operative hours.

Secondary outcomes: systemic vascular resistance index (SVRI), pulmonary vascular resistance index (PVRI), pulmonary blood flow (Qp), systemic blood flow (Qs), oxygen extraction ratio, and systemic oxygen delivery (DO2) were examined by measuring invasive mean arterial pressure (MAP), right atrial pressure (RAP), hemoglobin, arterial saturation by pulse oximetry, inspired oxygen, end-tidal carbon dioxide tension, oximetric SvO2 from SVC catheters, and assuming a pulmonary capillary saturation of 97%.

II. Are the results of the study valid?

Primary questions:

1. Was the assignment of patients to treatments randomized?

No, 62 patients received phenoxybenzamine, while 9 patients did not. It is unclear how these groups were assigned (e.g., were the 9 patients controls only because their families refused the study drug, or was there some other rationale for the assignment?).

2. Were all patients who entered the trial properly accounted for and attributed at its conclusion?

Was followup complete?

Obviously, no one was lost to follow-up during the 48-hour study period. 68 of 71 patients had data for the entire 48-hour period, unless they were supported on ECMO (One infant in each arm of the study received ECMO within the 1st 48 hours; oximetric data was excluded during ECMO). However, for 3 patients who had "early" deaths (all in the treatment group), it is unclear if any of these deaths occurred within the first 48 hours after surgery and thus there would not be data for the whole study period.

Were patients analyzed in the groups to which they were randomized?

Yes. There were no crossovers.

Secondary questions:

3. Were patients, health workers, and study personnel "blind" to treatment?

No. Again, the details are unclear, but it appears that study personnel and health workers were not blinded to treatment. As for whether or not the families knew, that is not clear. What can be inferred is that for the 2 patients whose parents did not consent for phenoxybenzamine, the families knew that their child did not get the drug.

4. Were the groups similar at the start of the trial?

Yes. There were no significant demographic or clinical differences between the two treatment groups. For example, the use of pre-operative mechanical ventilation and inotropic support was similar, as was aortic diameter. However it is interesting to note the trend toward longer CPB time among the control infants.

5. Aside from the experimental intervention, were the groups treated equally?

Between the two groups, generally, yes. All underwent the Norwood procedure with similar normalized Blalock-Taussig shunt sizes with similar anesthesia, cardiopulmonary bypass/modified ultrafiltration techniques, and milrinone infusions (during rewarming). However, vasoactive drug titrations were left to the discretion of the attending physicians (norepinephrine, nitroprusside, epinephrine) in the operating room to achieve an SVR of 12 WUm2.

Within the phenoxybenzamine group, in addition to all 62 patients receiving a fixed dose of phenoxybenzamine on CPB, 46 of 62 also received an infusion of phenoxybenzamine post-operatively to achieve a more consistent SVR reduction.

The control patients received higher doses of nitroprusside, dopamine, and epinephrine, while the phenoxybenzamine patients received higher doses of norepinephrine.

III. What were the results?

1. How large was the treatment effect?

In control patients, SvO2 peaked at an SaO2 range of 76-80% (with 77% giving the highest SvO2), with reduced SvO2 at SaO2 greater than 85% and at SaO2 less than 70%. The fitted fractional polynomial regression between SaO2 and SvO2 for control patients is as follows:

SvO2 = 49-0.064 x (77-SaO2)²    r² = 0.21, P < 0.0001

Given this relationship, the ΔCa-vO2 increased with SaO2 greater than 80% (P < 0.001).

In phenoxybenzamine patients, the SaO2-SvO2 relationship was continuously and linearly positive over the SaO2 range from 60-90%. The fitted fractional polynomial regression between SaO2 and SvO2 for this group is as follows:

SvO2 = 0.94 x SaO2 - 17    r² = 0.32, P < 0.0001

The ΔCa-vO2 was constant at all SaO2 (P = NS). The SvO2 was higher, and the ΔCa-vO2 lower across the entire SaO2 range with phenoxybenzamine (P < 0.0001), and the difference in SvO2 between groups increased progressively as SaO2 increased (P < 0.0001 by test for trend).

With respect to the other hemodynamic indices analyzed, the SVRI was lower in the treatment group (because phenoxybenzamine blocks alphaÐadrenergic vasoconstriction). At SaO2 greater than 80%, the following variables differed significantly between the treatment groups: SvO2, MAP, Qp:Qs, SVRI, DO2 (all with P < 0.0005), and PVRI (P < 0.05).

Phenoxybenzamine alters the SaO2-SvO2 relationship such that by reducing afterload, it allows for increased cardiac output (both to the pulmonary as well as to the systemic circulations). In the treatment patients, higher SaO2's (above 77-80%) reflected the increased cardiac output to both circulations. The increased pulmonary blood flow (resulting in increased SaO2) coupled with the increased systemic blood flow both contributed to an increased DO2. This in turn resulted in a decrease in the oxygen extraction ratio and therefore a linearly increasing SvO2 with increasing SaO2. On the other hand, in the control patients who lacked the phenoxybenzamine-mediated afterload reduction, a saturation above 77-80% reflected the relatively increased Qp (and thus SaO2) at the expense of decreased Qs. The reduction in systemic blood flow at high SaO2's exceeded the effects of increased SaO2's, thus leading to decreased DO2's, increased oxygen extraction ratios, and a lowering of SvO2's.

It is not entirely clear from the article how the 9 control patients were selected. The authors stated that since their earlier investigations showed better hemodynamics and improved survival with the use of phenoxybenzamine, they "have since administered phenoxybenzamine for all Norwood procedures, except for 2 patients without parental consent for phenoxybenzamine." Thus it seems like the majority of control patients were from an earlier time period. Although the demographic data did not demonstrate major differences between the two groups, the fact that most of the control patients were from an earlier time period introduces a bias. Perhaps the improved oxygen delivery in the treatment group was a result of better surgical techniques that come with more experience performing the Norwood procedure. Ideally, then, this study would have been better if it could have been randomized, with both groups undergoing the same procedure during the same era.

This study also showed that SaO2 is a poor predictor of SvO2. The sensitivity was 48%, the specificity 84%, and positive predictive value 5%. As can be seen from the r² values above, 21% and 32% of the SvO2 variance can be explained by SaO2 in the control and phenoxybenzamine groups, respectively.

IV. Will the results help me in caring for my patients?

1. Can the results be applied to my patient care?

This study focuses on one of the sickest patient populations in pediatric cardiac critical care. Using phenoxybenzamine allows clinicians to maintain higher SaO2's without compromising systemic oxygen delivery in a patient population at highest risk for low cardiac output syndrome. However, many centers now perform the Norwood procedure using the Sano modification (RV-PA conduit) instead of a Blalock-Taussig shunt to provide pulmonary blood flow (1). Sano shunts are usually 5-6 mm diameter Gore-tex tube grafts that are much longer than the 3-4.5 mm diameter B-T shunts. Since this study was performed on patients with B-T shunts, it is difficult to say whether patients with Sano shunts would respond the same way to afterload reduction using phenoxybenzamine. Most studies comparing the Sano modification with the B-T shunt have reported lower arterial saturations in Sano patients1. Does the SaO2/SvO2 curve depicted in Figure 3 maintain the same shape but merely shift to the left in Sano patients, or does the longer tube graft result in a different relationship altogether? This study needs to be repeated on Sano patients in order to answer this question.

This study is also important in emphasizing that one should not rely on SaO2 to predict SvO2. Instead, continuous direct SvO2 monitoring ought to be used to guide management of these very sick neonates post-operatively.

2. Were all clinically important outcomes considered?

No. This study focused primarily on hemodynamics during the first 48 hours after surgery. Although patients who received phenoxybenzamine did have improved systemic oxygen delivery, their mortality was higher overall. In the control group, 0/9 (0%) had early (survival to discharge after stage I palliation) deaths and 3/9 (33%) had late deaths. In the treatment group, 3/62 (5%) had early deaths and 14/62 (22%) had late deaths. The authors did not speculate on possible reasons for this difference. In fact, this study may be another example of how an intervention, despite apparent physiologic improvement, may have no effect on clinically meaningful outcomes.

3. Are the likely treatment benefits worth the potential harms and costs?

Phenoxybenzamine improved systemic oxygen delivery while maintaining higher SaO2's. In patients with atrioventricular valve leakage, afterload reduction can decrease the amount of regurgitation and improve cardiac output. Furthermore, a decrease in mean arterial pressure can help prevent bleeding and tearing of suture lines in the immediate post-operative period.

However, phenoxybenzamine has potential deleterious effects. As mentioned above, treatment with phenoxybenzamine increases the total cardiac output (to both circulations). While that is beneficial in the short term for preventing or minimizing low cardiac output, in the long run, increasing total cardiac output results in volume overloading of the heart. A systemic morphologic right ventricle pumping out a total of 8-9 L/min/m2 (including both Qp and Qs) could potentially lead to ventricular dysfunction over time. Perhaps this in part accounts for the increased mortality in the treatment group.

Also, phenoxybenzamine is a very long-acting vasodilator. Thus patients who are given this drug are committed to a treatment strategy based on sustained SVR reduction. This strategy is beneficial for the majority of patients, but in a vasodilated septic patient or in one who has a profound systemic inflammatory reaction to CPB, further pharmacologically-induced reduction of SVR in the face of an already low SVR can compromise cardiac output.

References:

1. Mair R, Tulzer G, Sames E, et al. Right ventricular to pulmonary artery conduit instead of modified Blalock-Taussig shunt improves postoperative hemodynamics in newborns after the Norwood operation. J Thorac Cardiovasc Surg. 2003 Nov;126(5):1378-84. [abstract]

Comments

Back to the EB Journal Club Index