Preoperative echocardiographic measures of left ventricular mechanics are associated with postoperative vasoactive support in preterm infants undergoing patent ductus arteriosus ligation

Abstract

Objective

Preoperative risk factors associated with poor outcomes after patent ductus arteriosus ligation in preterm infants have not been well defined. The aim of this study was to determine the association between preoperative echocardiographic measures of left ventricular mechanics and postoperative clinical outcomes after patent ductus arteriosus ligation.

Methods

Preterm infants less than 90 days of age with no other significant congenital anomalies who underwent patent ductus arteriosus ligation between 2007 and 2015 were considered for retrospective analysis. The primary outcome was peak postoperative vasoactive inotropic score. Conventional echocardiographic measures of ventricular size, function, and patent ductus arteriosus size were performed. Echocardiographic single-beat, pressure-volume loop analysis estimates of contractility (end-systolic elastance) and afterload (arterial elastance) were calculated. Ventriculoarterial coupling was assessed using the arterial elastance/end-systolic elastance ratio. Multivariable linear regression was performed using clinical and echocardiographic data.

Results

Echocardiograms from 101 patients (42.5% male) were analyzed. We found a statistically significant association between vasoactive inotropic score and both end-systolic elastance and arterial elastance. No patient with arterial elastance/end-systolic elastance greater than 0.78 (n = 32) had a vasoactive inotropic score 20 or greater. Analysis of our secondary outcomes found associations between preoperative end-systolic elastance and postoperative urine output less than 1 mL/kg/h at 24 hours, creatinine change greater than 0.5 mg/dL, and time to first extubation.

Conclusions

End-systolic elastance and arterial elastance were the only predictors of postoperative vasoactive inotropic score after patent ductus arteriosus ligation in preterm infants. Those neonates with increased contractility and low afterload were at highest risk for elevated inotropic support. These findings suggest a role for echocardiographic end-systolic elastance and arterial elastance in the preoperative assessment of preterm infants undergoing patent ductus arteriosus ligation.

Patent ductus arteriosus (PDA) is a frequently encountered form of congenital heart disease that increases the risk of significant morbidity and mortality in preterm neonates.^1–5 Surgical ligation is an established treatment for a hemodynamically significant PDA when a patient fails or is not a candidate for medical closure. In a subset of patients, PDA ligation leads to impaired left ventricular (LV) systolic performance and a “postligation syndrome” that has been described in the literature as oxygenation failure, hypotension, and the need for increased inotropic support.^1,2,6

The ability to preoperatively detect patients at high risk for postligation syndrome would allow clinicians to better anticipate the postoperative needs of their patients and ultimately improve care. Echocardiography is ideally suited to risk stratify these patients because the postligation syndrome is contributed to by an acute change in ventricular loading conditions leading to heart failure. Although previous studies have shown that postoperative echocardiographic surrogates of global LV systolic performance are predictive of cardiorespiratory instability after PDA ligation,³ no studies have evaluated the relationship between preoperative clinical findings and postoperative outcomes in these patients.

Echocardiographic estimates of contractility and afterload were recently validated in children against gold standard measures.⁴ The use of these simple measures is attractive because they allow us to gain insight into the contractile and loading states that predispose a preterm neonate to postligation syndrome and may ultimately be more powerful predictors of outcome than the conventional measures used in previous studies. The objective of this study was to determine the association between these preoperative novel echocardiographic measures of LV performance and postoperative outcomes after PDA ligation in preterm infants.

MATERIALS AND METHODS

This was a retrospective review of all infants undergoing surgical PDA ligation at our single institution between 2007 and 2015. The protocol was approved by our institutional review board.

Patient Selection

Patients were eligible if they had a gestational age less than 37 weeks and were identified if their surgical data were collected for inclusion in the Society of Thoracic Surgeons database. Infants were excluded from the study if they were (1) aged more than 3 months at the time of ligation, (2) had other congenital heart disease other than patent foramen ovale, (3) had other significant illness such as necrotizing enterocolitis or omphalocele, (4) had poor-quality images, or (5) had inadequate clinical data. Systematic chart review was performed on each patient to obtain detailed clinical data as described next.

Outcomes

The primary outcome was peak vasoactive inotropic score (VIS). The VIS is a validated measure of hemodynamic status based on inotropic need after pediatric cardiac surgery and has been validated as a predictor of prolonged intubation, intensive care unit stay, and total hospitalization in prior studies.⁵ The VIS is calculated as follows:

$$\mathit{VIS}\mathit{Score}=\mathit{dopamine}\mathit{dose}(\mu g/kg/\mathit{min})+\mathit{dobutamine}\mathit{dose}(\mu g/kg/\mathit{min})+[100\times \mathit{epinephrine}\mathit{dose}(\mu g/kg/\mathit{min})]+[10\times \mathit{milrinone}\mathit{dose}(\mu g/kg/\mathit{min})]+[10,000\times \mathit{vasopressin}(\mathit{units}/kg/\mathit{min})]+[100\times \mathit{norepinephrine}(\mu g/kg/\mathit{min})]$$

Vasoactive-inotropic score was calculated at 4, 8, 12, 24, 36, and 48 hours postoperatively. A patient was considered high risk if their peak VIS was 20 or greater.^7,8

Secondary outcomes included urine output less than 1 mL/kg/h at 24 hours, postoperative time to extubation, and increase in creatinine greater than 0.5 mg/dL. Creatinine levels were recorded preoperatively and at 24 and 48 hours postoperatively.

Echocardiographic Analysis

Preoperative echocardiograms were retrospectively analyzed from the clinical server (Xcelera, Philips Medical Systems, Andover, Mass). Measurements were recorded by a single blinded reviewer (M.A.G.). Echocardiographic timing varied between 0 and 14 days preoperatively. Noninvasive blood pressures were obtained at the time of the echocardiogram by automated sphygmomanometer. All echocardiographic variables were measured over 1 cardiac cycle.

PDA size was estimated using the PDA:left pulmonary artery (LPA) ratio the diameter of the LPA and PDA and were obtained in the ductal view (high left parasternal view) by measuring the internal diameter of the vessel during systole. Peak systolic PDA gradient was also recorded. Measures of LV size included end-diastolic and end-systolic volume calculated using the 5/6 * area * length method. The LV length was measured in both systole and diastole in the apical 4-chamber view. The LV area in cross-section was traced in both systole and diastole in the parasternal short-axis view. Ejection fraction was calculated as (end-diastolic volume – end-systolic volume)/end-diastolic volume. The LV internal dimension in diastole and fractional shortening was derived from M-mode.

Speckle-tracking echocardiography was performed to assess systolic myocardial deformation using Cardiac Performance Analysis v. 3.0 (Tomtec, Munich, Germany). Longitudinal strain and strain rate were measured by manually tracing from the lateral to the septal component of the mitral annulus in the apical 4-chamber view and averaging the resultant 6 segments. Circumferential strain and strain rate were measured in the parasternal short-axis view at the level of the papillary muscles. Images were excluded for analysis if more than 2 segments could not be adequately tracked.

End-systolic elastance (Ees) and arterial elastance (Ea) have been established as reference-standard measures of LV contractility and afterload, respectively. The ratio of these 2 (Ea/Ees) is a marker of ventriculoarterial coupling, an important component of global pump function.^9–11 These measures are derived from pressure-volume loop analysis and are generally not feasible to measure clinically because of the need for invasive catheterization.¹² To use these measures more routinely, single beat estimations of echocardiographic Ees and Ea (ie, estimates of Ees and Ea that do not require load alteration to be measured) have been developed and validated in children.^4,13 Ea was calculated as (0.9 × systolic blood pressure/stroke volume). Ees was calculated as (0.9 × systolic blood pressure/end-systolic volume).¹⁴ Ea and Ees were both indexed to body surface area to account for differences in body size between patients. Echocardiographic Ea/Ees was also calculated to assess ventriculo-arterial coupling.

Statistical Analyses

The Shapiro–Wilk test was used to determine whether data were normally distributed. Independent t tests or Mann–Whitney U tests were used to detect differences between groups. Differences between categoric groups were assessed using the chi-square or Fisher exact tests. All reported demographic, clinical, and echocardiographic characteristics (reported in Table 1) were considered for inclusion in the multivariable regression analyses. The associations between independent variables and VIS 20 or greater were evaluated using multivariable logistic regression. Variables were included in multivariable regression if the P value was less than .20 on univariate regression. Independent variables remained in the final models if they were statistically significantly associated with the dependent variable or their inclusion increased the explanatory power of the model significantly (increased R² ≥ 0.03). Multicollinearity was identified if the Variance Inflation Factor was greater than 10. If so, variables were removed from the model on the basis of decreasing effect on the explanatory power of the model. For this reason, ejection fraction was removed because of its high collinearity with Ea/Ees (variance inflation factor >10). Ejection fraction was noted to have a small impact on the explanatory model (R² went from 0.29 to 0.28). The associations between independent variables and urine output less than 1 mL/kg/h at 24 hours and peak creatinine increase of 0.5 mg/dL or greater were evaluated using multivariable logistic regression. Predictors of time to extubation were identified using multivariable linear regression. Receiver operating characteristic (ROC) curves were created and the area under the curve (AUC) was determined to assess the ability of each independent variable to predict the outcome. Statistics were performed using SPSS v. 23 (IBM, New York, NY).

TABLE 1.

Demographic, clinical, and echocardiographic data between preterm neonates with low versus high vasoactive inotropic score after patent ductus arteriosus ligation

Data	VIS <20 (n = 91)	VIS ≥20 (n = 10)	P value
Age (d)	28 (19, 49)	24 (19, 35)	.43
Male, n (%)	39 (43%)	4 (40%)	1.00
Weight (kg)	1.1 (0.83, 1.62)	0.88 (0.76, 3.10)	.52
Height (cm)	36 (32, 41)	33.5 (29, 36)	.17
BSA (m2)	0.11 (0.09, 0.14)	0.09 (0.07, 0.17)	.46
Hospital LOS (d)	78.5 (19, 126)	103.5 (61, 130)	.41
Postoperative LOS (d)	56 (13, 99)	75.5 (35, 119)	.33
Gestational age (wk)	25 (24, 27.5)	25 (23.8, 27.5)	.80
SBP (mm Hg)	68 ± 15	58 ± 17	.07
DBP (mm Hg)	33 (25, 42)	36 (23, 39)	.85
PDA:LPA	0.7 (0.55, 0.98)	0.82 (0.58, 0.99)	.62
LS (%)	−18.0 (−23.0, −14.5)	−20.8 (−25.3, −15.2)	.52
LSR (1/s)	−2.1 (−2.7, −1.5)	−2.1 (−2.8, −1.9)	.57
CS (%)	−15.4 (−19.5, −11.8)	−17.4 (−18.8, −15.8)	.20
CSR (1/s)	−1.8 (−2.2, −1.4)	−2.0 (−2.2, −1.7)	.32
PDA gradient (mm Hg)	20 (10, 33)	18 (13, 34)	.73
LVIDd (cm)	1.7 (1.3, 2.0)	1.3 (1.3, 1.7)	.09
FS %	0.41 (0.35, 0.44)	0.39 (0.36, 0.49)	.95
EDV (mL)	4.3 (2.8, 6.6)	2.7 (2.5, 4.6)	.11
EF %	0.60 ± 0.09	0.64 ± 0.05	.20
Ees (mm Hg/mL/m2)	3.6 (2.7, 5.1)	3.8 (2.9, 8.9)	.70
Ea (mm Hg/mL/m2)	2.5 (1.9, 3.2)	2.3 (1.6, 4.2)	.76
Ea:Ees	0.69 ± 0.27	0.57 ± 0.13	.03
Creatinine change (mg/dL)	0.0 (0.0, 0.1)	0.3 (0.2, 0.3)	<.01
Time to first extubation (d)	12 (5, 26)	6 (4, 41)	.85
UOP first 24 h (mL/kg/h)	3.7 (2.4, 5.7)	1.3 (0.4, 4.2)	.10

Data reported as a mean ± standard deviation or median (interquartile range). VIS, Vasoactive inotropic score; BSA, body surface area; LOS, length of stay; SBP, systolic blood pressure; DBP, diastolic blood pressure; PDA, patent ductus arteriosus; LPA, left pulmonary artery; LS, longitudinal strain; LSR, longitudinal strain rate; CS, circumferential strain; CSR, circumferential strain rate; LVIDd, left ventricular internal diameter in diastole; FS, fractional shortening; EDV, end-diastolic volume; EF, ejection fraction; Ees, end-systolic elastance; Ea, arterial elastance; UOP, urine output.

RESULTS

There were 157 preterm infants from 2007 to 2015 who underwent surgical PDA ligation. A total of 56 patients were excluded for the following indications: A total of 8 infants had significant other illness (necrotizing enterocolitis, omphalocele), 2 infants had no inpatient records available, 9 had inadequate echocardiographic images, 25 infants had other significant congenital heart disease, and 12 infants were aged more than 3 months at the time of operation (Figure 1). There were 101 patients included in the analysis. These were divided into 2 groups: VIS less than 20 (90 patients) and VIS 20 or greater (11 patients). The demographic, clinical, and echocardiographic data from these patients are presented in Table 1. Two patients died within 30 days of PDA ligation; 1 patient had Ees = 13.2 mm Hg/mL, Ea = 4.9 mm Hg/mL, and Ea/Ees of 0.37; and 1 patient had Ees = 6.2 mm Hg/mL, Ea = 3.7 mm Hg/mL, and Ea/Ees of 0.60.

FIGURE 1.

Visual representation of applied patient exclusion criteria.

Primary Outcome: Vasoactive Inotropic Score

No demographic data, clinical data, or measures of PDA size were associated with VIS 20 or greater. Speckle-tracking measures of myocardial deformation were not associated with VIS 20 or greater. The final multivariable logistic regression model is shown in Table 2 (Nagelkerke R² = 0.28). There were statistically significant associations with a VIS 20 or greater and both Ees and Ea. For every 0.1 mm Hg/mL/m² increase in Ees, patients had a 1.3 times increased odds of having a peak VIS 20 or greater, indicating that those patients with a higher Ees were associated with a higher VIS postoperatively. For every 0.1 mm Hg/mL/m² increase in Ea, patients had a 0.7 times decreased odds of having a peak VIS 20 or greater, indicating that those patients with a lower preoperative Ea were associated with a higher VIS postoperatively. The ROC analysis revealed an Ees AUC = 0.61, Ea AUC = 0.54, and Ea/Ees AUC = 0.62 to predict VIS 20 or greater. No patients with Ea/Ees 0.78 or greater (n = 32) had a peak VIS 20 or greater (32% sensitivity, 100% specificity).

TABLE 2.

Multivariable logistic regression to predict postoperative vasoactive inotropic score 20 or greater

Variable	B	SE	P valu
EDV	−0.8	0.4	.17
Ea/Ees	−32.2	17.4	.06
Ees	2.8	1.5	.04
Ea	−3.2	1.1	.04

SE, Standard error; EDV, end-diastolic volume; Ea, arterial elastance; Ees, end-systolic elastance.

Secondary Outcome: Urine Output

Multivariable logistic regression revealed statistically significant associations between urine output less than 1 mL/kg/h at 24 hours and age at operation (P = .02), PDA:LPA ratio (P = .03), ejection fraction (P = .04), Ea/Ees (P = .04), Ees (P = .02), longitudinal strain rate (P = .02), and end-diastolic volume (P = .04). The ROC analysis revealed an Ees AUC = 0.64.

Secondary Outcome: Creatinine Change

Multivariable logistic regression revealed statistically significant associations between creatinine change greater than 0.5 mg/dL and only Ees (B = 0.21, P = .04). The ROC analysis revealed an Ees AUC = 0.72.

Secondary Outcome: Time to Extubation

Multivariable linear regression revealed statistically significant associations between time to first extubation and weight (β = −0.58, P < .01) and Ees (β = 0.34, P<.01). The ROC analysis revealed an Ees AUC = 0.75. No patient with Ea/Ees greater than 0.75 had time to first extubation greater than 20 days. All ROC curves are available in Figures E1 to E4.

DISCUSSION

To our knowledge, this is the first study to show that preoperative echocardiographic measures of myocardial mechanics can be used to risk stratify preterm infants at high risk for morbidity after PDA ligation.

Echocardiographic Predictors of Morbidity After Patent Ductus Arteriosus Ligation

We found a positive association between Ees and VIS and a negative association between Ea and VIS. In addition, our results suggest that an Ea/Ees ratio 0.78 or greater may be useful in identifying low-risk patients, but its sensitivity is limited. Ees also had positive associations with the secondary outcomes, including urine output less than 1 mL/kg/h at 24 hours postoperatively, creatinine change greater than 0.5 mg/dL postoperatively, and time to first extubation. These results suggest that the assessment of preoperative myocardial mechanics may be clinically useful in preterm infants who are scheduled to undergo PDA ligation.

A few prior studies have evaluated clinical characteristics and conventional echocardiographic data to determine the risk of morbidity and mortality associated with PDA ligation.^15–17 However, previous studies have been limited by their omission of the assessment of myocardial function in their analyses. For example, one study found an association between preoperative PDA velocity and duration of mechanical ventilation postoperatively.¹⁸ This is in contrast to our results that showed no relationship between outcomes and PDA velocity. PDA velocity likely was not associated with outcomes in our results because we included measures of function in our analysis that are stronger predictors of outcome compared with PDA velocity. One study used echocardiographic measures of function to predict outcomes in this population. Jain and colleagues³ showed that a decline in LV output 1 hour postoperatively was associated with cardiorespiratory instability and the need for increased inotropic support. However, they did not include preoperative measures in their analysis, limiting their data’s use in preoperative risk stratification. This study, to our knowledge, is the first to show that preoperative echocardiographic measures of LV mechanics can be used as a predictor of postoperative outcomes after PDA ligation in preterm infants.

Insights Into the Role of Ventricular Mechanics

Our results suggest that patients with higher LV contractility and lower afterload on preoperative echocardiography had increased odds of having a higher VIS postoperatively. Initially, it may seem counterintuitive that the infants with higher contractility in the setting of a lower afterload had worse outcomes; however, the phenomenon of an increased contractility with decreased afterload leading to poor cardiac outcomes is present in other disease processes. Cantu syndrome is associated with a gain of function mutation in the K_ATP channel leading to decreased systemic vascular resistance and increased cardiac contractility.^19,20 These patients over time develop an increased cardiac output in the setting of chronic vasodilation that leads to an LV volume overload and dilation. This setting of a high cardiac output state leads to worsening cardiac function over time.²¹ A similar mechanism conceivably can be applied to the preterm PDA cohort. The infants with a larger PDA have a resultant lower afterload. When contractility is higher in this environment, the inappropriate ventriculoarterial interaction leads to a high output heart failure state. Dysfunctional ventriculoarterial interactions may respond poorly to the abrupt changes in loading conditions postoperatively, placing the infant at risk for cardiorespiratory instability.^1,22

Clinical Implications

The use of noninvasive novel echocardiographic measures of LV contractility and afterload in preterm infants with PDA has several clinical implications. Our data were able to determine which infants were at low risk for cardiorespiratory instability postoperatively. This information would be useful to families and caregivers making decisions regarding PDA ligation and risk–benefit analysis for medical team members. This information is valuable because untreated, hemodynamically significant PDA has been associated with significant morbidity and mortality,^23–26 but determining when or if to intervene has proven difficult.^27–29 If we can determine which infants are at low risk for postoperative morbidity, deciding if or when to surgically ligate would be more straightforward for the patient’s family and caregivers. For example, in patients with a hemodynamically significant PDA and an Ea/Ees greater than 0.78, there may be a lower threshold for undergoing PDA ligation as the risk for low cardiac output after surgery is low. For patients with a hemodynamically significant PDA and an Ea/Ees less than 0.78, there is some risk of requiring high inotropic support afterward. This should factor into the decision whether PDA ligation is warranted or not. Future studies should assess whether changes in timing of surgery or preoperative initiation of vasoactive support are warranted in patients with Ea/Ees less than 0.78.

Study Limitations

The main limitation of this study is inherent to the nature of a retrospective chart review. This type of study can only provide us with correlations and not causation. VIS has been validated in infants/children undergoing cardiac surgery that requires cardiopulmonary bypass, but not necessarily in closed heart operations. As such, the ability of VIS to predict outcomes in this population is unknown. Even so, we think this score was an appropriate method to assess acute inotropic need after PDA ligation, as have Davidson and colleagues.⁸ We were also limited to the data that were available to us in the patient chart and echocardiographic records. Data had to be omitted when it were not available in the chart or echocardiographic images were not clear enough for measurements. Of course, preoperative variables not accounted for in this study, such as preoperative inotropic support, may influence postoperative inotropic needs and may warrant future study. There was no consistency in the timing of the echocardiograms obtained with respect to when surgery occurred and the protocols used to obtain the images. All of the echocardiograms were obtained preoperatively, but varied between on the day of surgery to 2 weeks before surgery. In addition, all echocardiographic measurements were measured using 1 cardiac cycle because of limitations in the original acquisitions. This may increase measurement variability when compared with measures averaged over multiple beats. Last, although Ees and Ea have been validated in children, they have not been well studied in preterm infants or validated in the setting of acute changes in loading conditions or contractility.⁴

CONCLUSIONS

A preoperative echocardiographic measure of contractility, Ees was associated with peak VIS, urine output, creatinine change, and time to first extubation in preterm infants undergoing surgical ligation. Preterm infants were at low risk for high postoperative inotropic needs when Ea/Ees was greater than 0.78. These noninvasive echocardiographic measures of LV contractility, afterload, and ventriculoarterial coupling have potential clinical utility perioperatively in preterm infants undergoing surgical PDA ligation to assist with risk stratification and medical decision making.

Supplementary Material

FIGURE E1. ROC curve for VIS greater than 20. ROC, Receiver operating characteristic; VIS, vasoactive inotropic score; Ees, end-systolic elastance; Ea, arterial elastance.

FIGURE E2. ROC curve for time to extubation greater than 20 days. ROC, Receiver operating characteristic; Ees, end-systolic elastance; Ea, arterial elastance.

FIGURE E3. ROC curve for urine output less than 1 mL/kg/h at 24 hours. ROC, Receiver operating characteristic; Ees, end-systolic elastance; Ea, arterial elastance.

FIGURE E4. ROC curve for creatinine change greater than 0.5 mg/dL. ROC, Receiver operating characteristic; Ees, end-systolic elastance; Ea, arterial elastance.

Central Message.

Preoperative echocardiographic measurement of LV Ees and Ea may help predict postoperative change in clinical status in preterm infants undergoing PDA ligation.

Echocardiographic measures of LV mechanics in preterm infants undergoing PDA ligation.

Perspective.

PDA is a common form of congenital heart disease in preterm infants. Surgical ligation has been associated with cardiorespiratory instability. This study found an association between preoperative increased Ees and decreased Ea with a postoperative increase in vasoactive support after PDA ligation. These data are valuable in facilitating risk analysis and management.

Abbreviations and Acronyms

AUC

area under the curve

Ea

arterial elastance

Ees

end-systolic elastance

LPA

left pulmonary artery

LV

left ventricular

PDA

patent ductus arteriosus

ROC

receiver operating characteristic

VIS

vasoactive inotropic score