Case Scenario: Perioperative Management of a Young Woman with Fontan Repair for Major Gynecologic Surgery

CONTINUOUS improvement of surgical techniques over the last decades has allowed many patients with congenital heart disease to survive in good clinical conditions through adulthood. As such, more and more frequently, some of them need to recur to noncardiac surgery.¹  However, the perioperative management of these patients requires specific skills and appropriate knowledge of the preexisting cardiovascular conditions and the pathophysiologic effects of palliative interventions.²  Particularly, the Fontan procedure is a surgical technique that allows the final palliation of many congenital heart defects.³  The final result consists of anastomosis of both superior and inferior vena cavae to pulmonary artery with pulmonary veins flowing in a common atrium and a single ventricle⁴  (right or left depending on the initial cardiac pathology) providing blood flow to systemic circulation (fig. 1). A variable degree of right-to-left shunt may be present to counteract abrupt increase in pulmonary vascular resistance (PVR).⁵  In such condition, some important physiologic changes occur: pulmonary blood flow is entirely dependent on venous return (VR) owning to the absence of right ventricular pump, so that any increase in PVR determines a venous congestion with decreased cardiac output (CO). In this condition, the single ventricle may suffer from a different degree of both systolic and diastolic dysfunction, which can finally lead to tissue hypoperfusion. General anesthesia and positive pressure ventilation are likely to profoundly affect such modified physiology, with consequences that need to be carefully evaluated in the preoperative phase. However, an accurate intraoperative monitoring is advisable in such condition, especially in patients with failing Fontan and urgent surgery or who undergo major surgery. CO and cardiac filling pressure monitoring are of pivotal importance, but they are hardly achievable. Indeed, a standard pulmonary artery catheter cannot be used because of altered anatomy, and also central venous line placement is debatable, although very useful, because recorded central venous pressures (CVPs) are, indeed, mean pulmonary artery pressures (MPAPs), given there is no stenosis in the baffle or extraconduit. In this article, we will discuss the anesthetic management of a young patient with Fontan physiology who underwent major gynecologic surgery, whose hemodynamic monitoring was carried out using a transpulmonary dilution technique (EV 1000, Edwards Lifescience, USA).

A 31-yr-old women was scheduled for gynecologic surgery because of symptomatic endometriosis, which led the patient to severe anemia due to metrorrhagia (hemoglobin, 6 g/dl) with consequent acute cardiac failure in August 2014. When the patient was born, she was diagnosed with double outlet left ventricle with doubly committed ventricular septal defect and transposition of great arteries. She underwent multiple palliative surgeries, completed in 1989 with a definitive extracardiac nonfenestrated Fontan with subsequent substitution of the conduit between inferior vena cava and pulmonary artery in 2013 because of stenosis. After the episode of acute anemia in August 2014, her recent medical history did not yield any sign of decompensated heart failure. In the preoperative evaluation, performed by a team including a cardiologist, a gynecologist, an anesthesiologist, and a cardiac surgeon, the patient showed a good residual functional capacity estimated in 4 to 10 metabolic equivalents, she referred abdominal pain and metrorrhagia. She was classified as a New York Heart Association II to III (dyspnea for moderate efforts) without angina. The last ECG Holter reported an aritmic burden less than 1%. An echocardiography performed some days before surgery described nondilated ventricle with mild contractility impairment (ejection fraction, 53%), with normal flows on both superior and inferior vena cavae. At clinical examination, the patient was quite unremarkable, her weight was 45 kg, and she was 160-cm tall with a body surface area of 1.44 m². The night before surgery, a peripheral venous line was placed (18 gauge) to allow hydration by lactated Ringer’s solution overnight (42 ml/h). The surgery was considered at increased risk because of possible intraoperative unexpected complications with a duration ranging from a few to several hours. Accordingly we decided to perform hemodynamic monitoring using a transpulmonary dilution technique (EV 1000, Edwards Lifescience). After patient arrival in preanesthesia room, she was reassured and administered with 1 mg of iv midazolam. Then, she entered in the operating theater, and an arterial femoral line was placed after local anesthesia with 10 ml lidocaine, 2%, with an echo-guided technique. Electrocardiogram, peripheral Spo₂ and bispectral index were used as monitoring during surgery. A total of 1.5 mg midazolam, 1.5 mg/kg ketamine, and 3 μg/kg fentanyl were administered at induction. As soon as hypnosis was achieved (bispectral index, <60), 0.6 mg/kg rocuronium was administered, and endotracheal intubation was successfully accomplished with an endotracheal tube n.7. Mechanical ventilation was set as follows: 0 cm H₂O positive end-expiratory pressure (PEEP) and tidal volume 6 ml/kg with a respiratory rate ranging from 12 to 16/min, set to achieve an end-tidal CO₂ between 30 and 35 mmHg and a Fio₂ ranging from 0.3 to 0.4 to achieve a saturation more than 96%. Peak and plateau pressures during the surgery were kept below 14 and 12 cm H₂O, respectively. Anesthesia was maintained with sevoflurane (minimum alveolar concentration, 0.5 to 0.8) and remifentanil (0.05 to 0.3 μg kg⁻¹ min⁻¹). Despite 500 ml of lactated Ringer’s solution infused before induction, a slight hypotension occurred after induction with recovery of mean arterial pressure after infusion of 3 mg etilefrine and other 700 ml of crystalloids. Thereafter a central venous line was placed in internal jugular right vein with echo guide and was secured at 12 cm. No complications occurred during the procedure. CVP value was 18 mmHg, with 69% ScvO₂. The first hemodynamic measures recorded after induction are reported in table 1. To avoid any risk of air injection during thermodilution, we used a syringe prefilled with 20 ml of saline of which only 15 ml was injected, secured to venous line by a luer lock system. Duration of surgery was quite longer than scheduled, because after hysteroannessiectomy and appendicectomy, urologists have been involved because of need of reimplantation of ureter owing to stenosis caused by an endometriosis nodule. Thereafter the radiologist was also called to perform a nephrostomy, so that we left operating theater after 10 h since patient arrival, with a total time of anesthesia of 9 h and 30 min. Only one major episode of hemodynamic instability occurred with an abrupt decrease of mean arterial pressure (up to 58 mmHg) and CO, which were rapidly restored by fluids infusion (crystalloids and albumin). Diuresis monitoring was quite challenging during surgery, because surgeons needed to check continuously bladder integrity by injecting methylene blue through urinary catheter, so that we could estimate only at the end of surgery a total diuresis of 400 ml, with estimated blood loss of 300 ml and a total balance of +1,200 ml. Considered the risk of profound sympathetic block induced by epidural analgesia and, given the prolonged preplanned monitoring, postoperative analgesia was guaranteed by acetaminophen administration in an around-the-clock regimen and an elastomer, which delivered iv morphine and ketorolac at a rate of 0.6 and 3 mg/h, respectively, for 30 h. The patient was admitted to intensive care unit for postoperative monitoring and was successfully extubated after an hour from arrival. The following day, her hemodynamic parameters were unremarkable as well as her arterial blood gas analysis, so she was discharged from intensive care unit to gynecologic ward.

In this article, we reported the anesthetic management of a young patient with previous Fontan palliative surgery for congenital heart disease, who underwent a complex gynecologic surgery. Fontan palliation is adopted for several pathologic conditions to allow a series blood circulation, regardless of the absence of right ventricular pump.⁶  More specifically, the Fontan procedure consists of anastomosis of superior and inferior vena cavae with pulmonary artery. The latter may be done through an intraatrial tunnel or an extracardiac conduit, which, in either case, is referred of as “Fontan baffle.” Fenestration may be present on the conduit to allow the blood to flow to the right atrium in case of increased PVR, at the expense of right-to-left shunt, to maintain CO.⁷  This final anatomical situation implies some important physiologic changes that must be well kept in mind, particularly when we are going to deeply perturb this equilibrium with a general anesthesia (fig. 2).

In our body, each flow is determined by the difference between an upstream and a downstream pressure divided the resistances to the flow itself. When we refer to VR, upstream pressure is considered the pressure that stretches vessels and can be measured inducing a temporary circulatory standstill or extrapolated with inspiratory holds at increasing pressures in mechanically ventilated patients.⁸  This pressure is called mean systemic filling pressure (Pmsf). Accordingly downstream pressure is the pressure of the heart chamber where such flow is directed, that is, right atrium pressure (Pra). Hence, the determinants of VR may be described as follows:

where Rv are the resistances to VR.

In an adult healthy young man at rest (fig. 1A), Pra is close to 0 mmHg, Pmsf is close to 7 mmHg, and Rv are 1.4 mmHg l⁻¹ × min⁻¹ with a consequent VR of 5 l/min.⁹  Indeed such value corresponds exactly to CO, and it could not be otherwise because CO and VR must be identical, except for minimal discrepancies due to pleural pressure variations during respiratory cycle. After Fontan palliative surgery, downstream pressure experiments a rapid increase because there is no ventricular activity to keep the pressure low, any longer, so Pra turns, from a functional point of view, into MPAP; however, we will refer to Pra in such context as CVP to avoid any confusion. Accordingly, Pmsf must increase to preserve the gradient for VR, up to very high values (fig. 1B). In our case, if we consider hemodynamic values at surgical incision and assume that Rv were stable at 1.4 mmHg l⁻¹ × min⁻¹ to obtain a VR of 3.6 l/min with a CVP of 20 mmHg, Pmsf must be at least 25 mmHg. Pmsf is determined by the amount of blood stretching the vessels, namely stressed volume, and by the venous tone. Anesthetic drugs induce a decrease in vascular tone, including venous tone that is crucial in this case and cardiac contractility that may lead to detrimental consequences. Hypotension observed in our patient after anesthesia induction was indeed promptly corrected by fluid infusion that rapidly increases Pmsf and VR. A venoconstrictor such as phenylephrine is also likely to be effective in increasing Pmsf by reducing unstressed volume in favor of stressed volume. However, a vasoconstrictor-like phenylephrine may also have negative effects. On one hand, it may induce a slight increase in Rv that may counterbalance the increase in Pmsf without any significant amelioration of preload. Conversely, the induced increase in afterload may worsen ventricular function, especially in the setting of preexisting systolic dysfunction. Also positive pressure ventilation importantly affects hemodynamic status, owing to augmented intrathoracic pressures,¹⁰  leading to both increased resistance to VR and increase of PVR because of alveolar overdistension. In this context, the use of low PEEP may be important to prevent collapse of alveoli and consequent increase in PVR, but excessive level of PEEP may also contribute to overdistension and increase of PVR, so that end-expiratory pressure must be carefully adjusted based on the patient’s specific characteristics. In the present case, to minimize the impact of mechanical ventilation, we decided to apply 0 cm H₂O of PEEP and low-tidal volumes with higher respiratory rate to keep a low end-tidal CO₂, to avoid hypercapnia, which may aggravate pulmonary hypertension by means of increased PVR.

However, transpulmonary pressure gradient also must be considered as shown in equation 2

where MPAP is mean pulmonary artery pressure, PAOP is pulmonary artery occlusion pressure, and Rp are pulmonary resistance. In patients with Fontan physiology, MPAP corresponds to CVP because no right ventricular activity is present. Also, CO can be substituted by VR, as we aforementioned, so we can rearrange equations 1 and 2 to investigate the determinants of CVP as shown in equation 3

From such equation, we can make some further consideration. First, in patients with Fontan physiology, an increase in either Pmsf or pulmonary artery occlusion pressure or both will determine an increase in CVP and vice versa, albeit pathophysiologic backgrounds are completely different. Second, an increase in Rp leads CVP to increase, whereas an increase in Rv implies a decrease in CVP. Third, and maybe most important, such equation underlines that a pressure gradient among Pmsf, CVP, and pulmonary artery occlusion pressure must always be preserved to assure an adequate blood flow. Otherwise, blood flow drops and congestion rapidly occurs. Table 2 briefly summarizes some strategies to maintain an adequate CO and keep Rp low.

Such complexity makes the correct interpretation of CVP in this category of patients quite challenging, particularly in the absence of other data concerning CO and/or volemic status. In addition, dynamic indexes of fluid responsiveness, namely pulse pressure variation and stroke volume variation,¹¹  may be not reliable in such scenario. Indeed, on one hand, they have never been tested in patients with such altered anatomy and physiology, and, conversely, the use of low tidal volumes would suggest a lower threshold to define fluid responsiveness, but this issue is still under debate.¹²  Consequently, to obtain a more complete monitoring, we decided to adopt a transpulmonary thermodilution (TPTD) methodology during anesthesia (EV 1000, Edwards Lifescience). Such technique allows direct measure of CO with intermittent TPTD and continuous calibrated CO monitoring through pulse contour methodology. The device requires a femoral arterial line placement, which measures stroke volume and temperature variation over time during the thermodilution through a thermistor on the tip of arterial line, along with a central venous line to inject cold fluid bolus. In this case, we decided to insert a central venous catheter allowing continuous measurement of ScvO₂ as well (PreSep, Edwards Lifescience). TPTD and pulse contour provide a more accurate measure of CO, especially in situations in which rapid changes in arterial elastance are expected, compared with noncalibrated methodology.^13,14  Furthermore, TPTD provides important information about intrathoracic volumes, which we thought being helpful for a correct intraoperative management. However, some potential pitfalls of such technique in this category of patients must be acknowledged. TPTD calculates intrathoracic volume through the analysis of thermodilution curve, more precisely through mean transit time, i.e., the time required for half of the indicator to pass the thermistor in femoral artery, and exponential downslope time of thermodilution curve. Patients with Fontan palliation are likely to have a certain degree of shunt (right-to-left or left-to-right shunt) due to incomplete pulmonary artery ligation, coronary sinus blood return in common atrium, or other preexisting shunts from pulmonary vessels.²  In case of left-to-right shunt, the aforementioned time intervals result slightly increased, particularly downslope time results increased up to 50%.¹⁵  As a consequence, global end-diastolic volume index (GEDVI) is likely to be slightly underestimated and extravascular lung water index (EVLWI) is overestimated (see tables 2 and 3). Although we adopted EV1000 technology (Edwards Lifesciences, USA), which uses a slightly different formula to calculate GEDV (see table 2) theoretically less sensitive to recirculation, an overestimation of EVLWI and underestimation of GEDV were likely observed. Indeed, we recorded low GEDVI and high EVLWI during our monitoring, in the absence of any clinical signs of either pulmonary congestion or gas exchange impairment. So we chose not to rely on absolute GEDVI value but rather to follow the changes of its value over time and after therapeutic intervention to assess the appropriateness of our management, because of this limitation in such specific category of patients.

Transoesphageal echocardiography (TEE) would also be very helpful to increase accuracy of hemodynamic monitoring in such patients during prolonged surgery. Indeed, TEE allows both detection of small, but potentially catastrophic, air embolism due to accidental abdominal vein laceration during surgery and shunt detection by comparing right and left stroke volumes. In addition, TEE provides information regarding both systolic and diastolic function and possible valve dysfunctions (i.e., stenosis and regurgitation). Unfortunately, it should be always integrated with a beat-to-beat continuous monitoring, because it cannot be performed continuously.

Finally, continuous ScvO₂ monitoring was quite useful as another marker of tissue hypoperfusion. ScvO₂ reaches lower values when O₂ peripheral extraction is increased, that is, when CO and/or hemoglobin concentration are low, so that an imbalance between oxygen delivery and consumption occurs. Low values recorded since surgical incision were compatible with a situation of low CO as testified by TPTD findings. However, despite values of ScvO₂ ranging between 65 and 69%, we decided to have lower target as a trigger to adopt interventions aiming at increasing oxygen delivery (DO₂). Indeed, because of changed anatomy due to Fontan palliation, in our patient, ScvO₂ might be considered more properly as SvO₂ (mixed O₂). Also, we decided to integrate venoarterial carbon dioxide gradient (_v-aCO₂ gap), which is a marker of CO inadequacy when it is more than 6 mmHg,¹⁶  in the decisional tree along with other hemodynamic variables. So, only during the hypotensive episode occurring toward the end of surgery, we corrected rapidly CO by means of fluid administration, owing to a clinical evidence of low preload (drop in CVP) and low CO with tissue hypoperfusion (ScvO₂ < 50%, _v-aCO₂ gap >6 mmHg). Interestingly in such situation, no increase in blood lactate was observed (table 1), probably due to the more rapid response of _v-aCO₂ gap after peripheral hypoperfusion, which helped us to correct it promptly, avoiding any development of lactic acidosis.

Perioperative management of patients with Fontan palliation, who undergo major surgery, is quite challenging for the anesthesiologist. However, a deep comprehension of the altered anatomy and physiology of these patients, along with an appropriate hemodynamic monitoring, may help the clinician to carry out the surgery with hemodynamic stability and without major tissue hypoperfusion. A deep knowledge of hemodynamic monitoring, including their potential pitfalls, is also indispensable to adopt wise therapeutic choices.

Support was provided solely from institutional and/or departmental sources.

The authors declare no competing interests.