Catecholamine-resistant hypotension and myocardial performance following patent ductus arteriosus ligation

Abstract

Objective

We performed a multicenter study of preterm infants, who were about to undergo patent ductus arteriosus ligation, to determine if echocardiographic indices of impaired myocardial performance were associated with subsequent development of catecholamine-resistant hypotension following ligation.

Study Design

A standardized treatment approach for hypotension was followed at each center. Infants were considered to have catecholamine-resistant hypotension if their dopamine infusion was >15 µg/kg/min. Echocardiograms and cortisol measurements were obtained between 6 and 14 hours after the ligation (prior to the presence of catecholamine-resistant hypotension).

Results

45 infants were enrolled: 10 received catecholamines (6 were catecholamine-responsive, 4 developed catecholamine-resistant hypotension). Catecholamine-resistant hypotension was not associated with decreased preload, shortening fraction or ventricular output. Infants with catecholamine-resistant hypotension had significantly lower levels of systemic vascular resistance and postoperative cortisol concentration.

Conclusion

We speculate that low cortisol levels and impaired vascular tone may play a more important role than impaired cardiac performance in post-ligation catecholamine-resistant hypotension.

Introduction

Approximately 30% of preterm infants who undergo surgical ligation of the patent ductus arteriosus (PDA) develop systemic hypotension and receive catecholamine support during the first 6-to-24 hrs after the procedure (1–4). Although catecholamine infusions can usually correct the post-ligation hypotension, approximately half of the hypotensive infants develop catecholamine-resistant hypotension (1, 4). In these infants, a "low stress-dose" of hydrocortisone can normalize their blood pressure (3, 5, 6). Epidemiologic studies have shown that decreased gestational and postnatal age are the strongest predictors of the postoperative hypotension (1–3); in contrast, differences in surgical, anesthetic, and intraoperative fluid management appear to have little effect on the post-surgical hemodynamic decompensation (1, 4, 7).

Changes in myocardial loading conditions and impaired myocardial performance may contribute to the postoperative clinical deterioration since increases in afterload and decreases in preload are observed immediately following PDA ligation (5, 6, 8–11). Although measurements of myocardial performance, made shortly (1–2 hours) after ligation, have not been found to have a strong association with the development of post-ligation hypotension (5, 10, 12), indices of impaired myocardial performance, at 6–8 hours after the operation (closer to the time of clinical deterioration), appear to have a much stronger relationship (6). Prophylactic treatment with milrinone, starting shortly after ligation, has been reported to reduce the incidence of cardiorespiratory instability; however, despite early milrinone treatment, some infants continue to develop post-ligation catecholamine-resistant hypotension (8, 9). To date, no studies have examined the relationship between indices of myocardial performance during the postoperative period and the incidence of catecholamine-resistant hypotension.

We recently conducted a prospective, multicenter study to determine the relationship between an infant's postoperative cortisol response and the development of catecholamine-resistant hypotension (4). We found that infants who developed catecholamine-resistant hypotension had significantly lower cortisol concentrations than infants who were normotensive or had hypotension responsive to catecholamine infusions and/or volume resuscitation. As part of the parent study, we performed an ancillary study to determine if (A) the development of catecholamine-resistant hypotension and (B) the presence of low postoperative cortisol concentrations were associated with indices of impaired myocardial performance. We report the results of this ancillary echocardiographic study below.

Methods

Eight of the original 12 study sites participating in the main study (4) had the resources to perform the ancillary study. After Institutional Review Board approval, 45 infants were enrolled following parental consent.

Inclusion/exclusion criteria, surgical, and postoperative management, as well as other details of the parent study can be found in our previous report (4). Infants were eligible for the study if they were 1) delivered between 23^0/7 – 31^6/7 weeks gestation, and 2) were about to have a PDA ligation. Although all infants had either a moderate or large left-to-right PDA shunt, there was no attempt to standardize the criteria used to determine the need for ligation (4).

Since catecholamine-resistant hypotension during the postoperative period was the primary outcome (see below), a standardized approach to management was agreed upon by all participating centers. The standardized approach outlined when volume expanders and vasopressors would be initiated, and the rate at which they would be increased. An arterial line and transducer were used to measure blood pressure continuously in all infants receiving catecholamine infusions or hydrocortisone for blood pressure support.

Hypotension was defined as "mean BP less than the 3^rd percentile for postmenstrual age (13, 14) that lasted more than 15 minutes". Operationally this meant that infants were considered to be hypotensive, and require treatment for their hypotension, if their mean blood pressure was less than [(postmenstrual age in mm Hg) − (3 to 4 mm Hg)]. When infants failed to maintain an adequate BP (defined as "BP greater than the hypotensive range"), 2–3 boluses of isotonic saline (10 mL/kg per bolus) could be given to correct presumed hypovolemia, before catecholamine support was initiated. Infusion of dopamine or dobutamine (the choice was left to the clinical neonatologist) was started at 5 µg/kg/min and increased by 2.5 µg/kg/min every 15 minutes until an adequate BP was achieved. Combinations of dopamine and dobutamine could be used. Milrinone was not used to treat post-ligation hypotension (however, at one study site (7 infants), a milrinone infusion was started shortly after ligation as part of an institutional protocol designed to prevent postoperative hypotension (8, 9).

If a dopamine infusion >15 µg/kg/min failed to maintain an adequate BP, a “low stress dose” of intravenous hydrocortisone (starting at 1 mg/kg/day) could be added. The “low stress dose” of hydrocortisone could not be used unless the dopamine infusion was >15 µg/kg/min. The collective experience, shared by the centers prior to the study, was that infants who failed to maintain an adequate BP, with dopamine infusions =15 µg/kg/min, usually required dopamine infusions >20 µg/kg/min before adequate BPs were achieved. Therefore, we classified infants as having Catecholamine-resistant hypotension if they received a dopamine infusion >15 µg/kg/min. In our study, all of the infants, who failed to maintain an adequate BP with dopamine infusions =15 µg/kg/min, were started on hydrocortisone (in addition to increasing the amount or number of catecholamine infusions).

We used a Catecholamine Score (calculated as the sum of all catecholamines corrected for potency: 1 × dopamine (µg/kg/min) + 1 × dobutamine) to measure the amount of catecholamine support an infant was receiving to maintain an adequate BP (15). A Vasoactive Inotropic Score (Catecholamine Score + 10 × milrinone (µg/kg/min)) was also calculated for each infant. (16).

We defined the severity of the hypotension by the maximum support needed to maintain an infant's BP above the hypotensive range: mild = volume boluses alone; Catecholamine-responsive hypotension = maximum Inotrope Score less than or equal to 15; and Catecholamine-resistant hypotension = maximum dopamine infusion >15 µg/kg/min.

Echocardiograms

We planned to perform our echocardiographic studies during the early stages of post-ligation hypotension, prior to the onset of Catecholamine-resistant hypotension (if it occurred). The decision to perform the study echocardiogram at this time point was based on several considerations. Prior studies have shown that echocardiograms performed either prior to, or 60 minutes after the surgery, are not associated with the development of post-ligation hypotension (5, 10, 12). On the other hand, echocardiographic measurements made at 6–8 hours after the operation (6) (when infants have already started to develop hypotension) have a strong relationship to the presence of postoperative hypotension. Unfortunately, the interpretation of measurements made in the presence of hypotension can be confounded by the need for and administration of inotropic drugs during the echocardiographic study. Since our goal was to determine if there was a relationship between measurements of impaired myocardial performance and the development of catecholamine-resistant hypotension, we planned to perform our studies between 6 and 14 hours after the ligation, at a time when hypotension was likely to be present, but before the onset of catecholamine-resistant hypotension.

Two-dimensional echocardiograms were performed by trained operators at each study site. All of the de-identified data recordings were analyzed by two of the authors who were unaware of the infants' degree of catecholamine support (SN, PM). The following echocardiographic measurements were made as previously described (5, 6): left ventricular output (LVO), stroke volume (SV), shortening fraction (SF), heart rate-corrected velocity of circumferential fiber shortening (VCFc), and LV end-diastolic diameter (LVEDD). Mean arterial BP (recorded at the time of the echocardiogram) was used to calculate wall stress (WS). We used mean BP because it is more representative of the entire cardiac cycle, and previous studies have shown an excellent correlation between mean arterial pressure and end-systolic pressure (17). Mean BP was also used to calculate systemic vascular resistance (SVR). We calculated the myocardial performance index (MPI), a widely used measure of global myocardial function, by dividing the sum of isovolumic contraction and relaxation times with the ejection time (18). MPI is inversely related to myocardial function. We were especially interested in indicators of impaired left ventricle systolic performance (SF ≤25% and LVO <150 ml/kg/min) (19, 20) since these have previously been reported to be associated with post-ligation hypotension and need for inotropic support (6).

Steroid measurements

Echocardiographic measurements were compared with serum cortisol measurements made between 8–14 hours after the surgery. If hydrocortisone was started before the 8–14 hour time point, the cortisol measurement was obtained before the first hydrocortisone dose was administered. Cortisol assays were performed by ultra-performance liquid chromatography tandem mass spectrometry (4, 21). Values obtained by this technique have correlation coefficients between 0.7 and 0.99 when compared with those obtained by radioimmunoassy (22). The cortisol results were not available during the study and were not used in the decision to start hydrocortisone treatment.

Since the adequacy of an infant’s cortisol response for maintaining blood pressure depends on both the circulating cortisol level and the stress experienced by the infant, we defined a “decreased cortisol response” as one in which the cortisol level was in the lower quintile of the postoperative cortisol concentrations found in the study population. By using this definition we could match the appropriateness of the cortisol response to the level of stress that the infant was experiencing since all infants underwent the same operation.

Statistics

Data were analyzed using Student’s t-test, Chi-Square or Fisher’s Exact, and the Mann-Whitney test as appropriate. Data are reported as mean ± standard deviation or median (interquartile range).

Results

Forty-five infants were enrolled in the ancillary echocardiographic study. Mean gestational age, birth weight, and postnatal age at ligation were 25.5±1.6 weeks, 782±190 grams, and 3.5±1.9 weeks, respectively. Forty percent (18/45) met criteria for hypotension: 8 were treated with volume boluses alone; 10 received catecholamines (dopamine, with or without dobutamine) in addition to volume boluses. Six of the 10 infants who received catecholamines were Catecholamine-responsive (normalized their blood pressure with Catecholamine Scores ≤15); 4 developed catecholamine-resistant hypotension (with dopamine infusion >15 µg/kg/min) and received hydrocortisone plus catecholamine infusions to normalize their blood pressure. As we observed in the parent study (4), infants in the ancillary study, who developed post-ligation hypotension (n=18), were more likely to be younger and weigh less at the time of ligation (27.9±2.6 versus 29.5±2.2 weeks post-menstrual age, p<0.04; 2.6±1.6 versus 4.1±1.9 weeks post-natal age, p<0.02; and, 910±338 versus 1162±353 grams, p<0.02, respectively) (1, 3).

Infants who received catecholamines to maintain an adequate BP, had them started within the first 12 hours after ligation. Maximal Catecholamine Scores were reached by 18 hours after ligation (Table 1). Among the infants who developed Catecholamine-resistant hypotension, Catecholamine Scores >15 did not occur until at least 10 hours after ligation (Table 1). All of the echocardiograms were performed prior to the presence of catecholamine-resistant hypotension and before maximal Catecholamine Scores were reached (Table 1).

Table 1.

Relationship between study measurements and the amount and timing of catecholamine support

	Normotension	Hypotension
	None (n=27)	Volume- boluses alone (n=8)	Catecholamine- responsive (n=6)	Catecholamine- resistant (n=4)
Time after ligation when:
Study Echocardiogram performed (hr)	10.0±5.0	9.5±3.1	11.1±4.5	7.4±1.5
Serum Cortisol measured (hr)	12.2±3.4	10.3±2.2	10.9±1.5	10.1±1.2
Catecholamine infusion started (hr)	-	-	6.0±5.0	2.5±2.4
Catecholamine Score >15 first reached (hr)	-	-	-	11.8±1.7
Maximum Catecholamine Score reached (hr)	-	-	13.3±5.1	14.8±3.0
Rate of catecholamine/inotropic administration at the time of the Echocardiogram
Catecholamine Score	-	-	3.0±2.3	10±0
Vasoactive Inotropic Score	0.5±1.1	1.5±1.6	3.6±2.9	10±0

Definitions: Volume-boluses, infants who received volume-resuscitation/boluses alone; Catecholamine-responsive Hypotension, infants whose maximum Catecholamine Score was ≤15; Catecholamine-resistant Hypotension, infants whose maximum dopamine infusion was >15 µg/kg/min.

Infants who developed post-ligation hypotension were more likely than normotensive infants to have echocardiographic measurements consistent with decreased myocardial performance (Table 2-Part A): hypotensive infants were more likely to have shortening fractions less than or equal to 25%, and, there was a trend for LVO to be lower and MPI to be higher in the hypotensive group as well (Table 2-Part A).

Table 2.

Cardiac function, hemodynamics, and serum cortisol in study infants (n=45)

	Part A		Part B
	Normotension	Hypotension	Hypotension
	No Hypotension (n=27)	Hypotension (all types) (n=18)	Volume- boluses alone (n=8)	Catecholamine- responsive (n=6)	Catecholamine- resistant (n=4)
LVO (ml/kg/min)	240±58	216±79§	201±40	224±132	235±38
LVO <150 ml/kg/min (%)	8	17	13	33	0
Shortening Fraction ≤25% (%)	15	50*	63	33	50
MPI	0.42±0.12	0.47±0.21	0.47±0.12	0.41±0.23	0.58±0.30
MPI >2z (%)	11	33§	25	33	50
LVEDD (cm/kg)	1.32±0.37	1.36±0.31	1.42±0.28	1.29±0.41	1.33±0.25
LA/Ao	1.51±0.34	1.44±0.53	1.54±0.49	1.32±0.33	1.39±1.02
SVR (mm Hg/L/kg/min)	167±74	187±99	202±78	216±132	110±44**
Heart Rate (beats/min)	155±13	162±19	153±14	160±24	180±5**
Stroke Volume (ml/kg)	1.56±0.42	1.34±0.46	1.33±0.33	1.36±0.72	1.31±0.24
VCFc (Circ/sec)	1.24±0.30	1.25±0.36	1.18±0.39	1.33±0.22	1.25±0.51
Wall Stress (g/cm2)	31±19	29±19	41±23	20±5	17±7§§
Cortisol (ng/ml) median (interquartile range)	24 (10–54)	42 (10–84)	60 (22–100)	51 (33–367)	8 (2–12)**

Definitions: Hypotension (all types), all hypotension groups combined (i.e., infants who received volume-resuscitation/boluses alone, catecholamine-responsive infants, and catecholamine-resistant infants).

LVO, left ventricular output; MPI, left ventricular myocardial performance index; MPI >2z, MPI greater than 2 standard deviations above the normative mean; LVEDD, left ventricular end-diastolic diameter; LA/Ao, ratio of left atrium to aortic root dimension; SVR, systemic vascular resistance; VCFc, velocity of circumferential fiber shortening (corrected for heart rate);

Values represent mean ± standard deviation or percent (%);

^*p-value <0.05,

^§<0.15: For comparisons between infants with No Hypotension and infants with Hypotension (all types)

^**p-value <0.05,

^§§<0.15: For comparisons between infants with Catecholamine-resistant hypotension and infants with milder types of Hypotension (Volume-resuscitation/boluses-alone and Catecholamine-responsive infants)

The goal of our study was to determine if echocardiographic indices of myocardial performance were more impaired in infants who developed catecholamine-resistant hypotension than in those who developed milder forms of hypotension (i.e., those with volume-bolus responsive or catecholamine-responsive hypotension). We found that most of the myocardial and cardiac output measurements were similar between the two groups of hypotensive infants (Table 2-Part B). There were, however, significant differences between the two groups in the magnitude of their heart rates and systemic vascular resistances (Table 2-Part B). There was also a trend for wall stress to be lower in the catecholamine-resistant hypotension group (Table 2-Part B).

Infants who developed catecholamine-resistant hypotension had lower postoperative cortisol concentrations than infants who developed milder forms of hypotension (i.e., hypotension that responded to catecholamine infusions and/or volume boluses) (Table 2-Part B). We defined a “decreased cortisol level” as one in which the post-ligation cortisol level was in the lower quintile of the distribution of postoperative cortisol concentrations. Seventy-five percent of the infants who developed catecholamine-resistant hypotension had cortisol concentrations in the lower quintile compared with only 12% of those who never developed catecholamine-resistant hypotension (p<0.02). Infants who developed catecholamine-resistant hypotension were younger and weighed less at the time of ligation. However, as we previously observed in the parent study, low cortisol concentrations were a significant and independent risk factor for developing catecholamine-resistant hypotension even when demographic differences were included in a multivariable model (4).

We compared infants with postoperative cortisol concentrations in the lower quintile with infants whose cortisol concentrations were above the bottom quintile (Table 3). We found that the two groups differed only in indices of afterload (wall stress) and heart rate (Table 3).

Table 3.

Cardiac Function and Hemodynamics in infants with Low Post-ligation Cortisol concentrations

	Post-ligation Serum Cortisol concentrations in Lower Quintile
	No (n=36)	Yes (n=9)
LVO (ml/kg/min)	224±71	251±54
LVO <150 ml/kg/min (%)	15	0
Shortening Fraction <25% (%)	25	11
MPI	0.43±0.14	0.43±0.24
MPI >2z (%)	20	22
LVEDD (cm/kg)	1.37±0.32	1.21±0.40
LA/Ao	1.53±0.43	1.29±0.36
SVR (mm Hg/L/kg/min)	183±89	146±64
Heart Rate (beats/min)	154±15	172±12*
Stroke Volume (ml/kg)	1.46±0.46	1.47±0.37
VCFc (Circ/sec)	1.22±0.32	1.32±0.34
Wall Stress (g/cm2)	33±19	17±6*

Definitions: see Table 2

Values represent mean ± standard deviation or percent (%).

^*p-value <0.05,

^§<0.15

Discussion

Prior studies have noted that myocardial performance appears to be impaired (decreased LVO, decreased SF, and increased MPI) following PDA ligation (5, 6, 8, 9). Our findings are consistent with the prior reports, and support the hypothesis that impaired myocardial performance may contribute to the development of postoperative hypotension. We found a significant association between decreased shortening fraction and the development of post-ligation hypotension. Similarly, there was a trend for hypotension to be associated with decreased LVO and increased MPI (Table 2-Part A).

The goal of our study was to determine if indices of impaired myocardial performance were also associated with a specific subset of hypotensive infants - those whose hypotension is refractory to catecholamine treatment. Our findings do not support an association between impaired myocardial performance and catecholamine-resistant hypotension (Table 2-Part B). Nor do they support a relationship between indices of impaired performance and the presence of decreased postoperative cortisol levels (which is associated with catecholamine-resistant hypotension) (Table 3). Infants in our study, who had decreased postoperative cortisol levels and catecholamine-resistant hypotension, had lower levels of systemic vascular resistance and afterload (Table 2-Part B and Table 3). These findings are more consistent with the hypothesis that low cortisol levels and catecholamine-resistant hypotension may be related to impaired vascular tone rather than to impaired cardiac performance. The fact that hydrocortisone treatment of preterm infants with catecholamine-resistant hypotension increases blood pressure by increasing systemic vascular resistance and afterload/wall stress (without altering cardiac output) is consistent with our findings and supports our hypothesis (23). We speculate that low postoperative cortisol concentrations may lead to down-regulation of vascular adrenergic receptors, which, in turn, may contribute to the failure of catecholamine infusions to maintain vascular tone and blood pressure (24, 25).

Our study has some important limitations:

In addition to the limitations inherent in the use of functional echocardiography in preterm infants, echocardiograms in our study were acquired by different sonographers - at different sites. However, the echocardiograms were analyzed by the same two authors who were unaware of the infants’ need for catecholamine support.

In our study there were low numbers of infants who developed catecholamine-resistant hypotension. This is not an unexpected finding since only 8–15% of preterm infants undergoing PDA ligation develop this condition (1, 5, 6). However, the low numbers may have underpowered the study to determine a true difference in myocardial performance.

The study was designed so the echocardiograms would be performed close to the time the infants were expected to develop catecholamine-resistant hypotension. However, this meant that 80% of the infants who received catecholamines during the postoperative period were already receiving low-dose catecholamine infusions at the time of their echocardiogram. Although the level of catecholamine support during the echocardiogram varied between patients, all of the patients were studied before their maximum Catecholamine Score was reached and prior to the development of catecholamine-resistant hypotension. Nevertheless, the variation in inotropic support at the time of the echocardiogram may have confounded our analyses and altered our ability to detect significant differences in myocardial performance between the groups.

Despite the small sample size and variation in inotropic support at the time of the echocardiogram, we did observe a significant difference between the groups in measures of systemic vascular resistance and afterload. We speculate that following PDA ligation, impairment of myocardial performance may contribute to the development of postoperative hypotension. However, once hypotension does occur, low cortisol concentrations and abnormal vascular tone may play a more important role than impaired cardiac performance in the development of refractory catecholamine-resistant hypotension. Only a randomized controlled trial, that examines the utility of cortisol replacement in hypotensive infants with low serum levels, will be able to elucidate the true contribution of low cortisol and low vascular tone on the severity of post-ligation hypotension.

Abbreviations

PDA

Patent ductus arteriosus

BP

Blood pressure

LVO

Left ventricular output

SF

Shortening fraction

SV

Stroke volume

VCFc

velocity of circumferential fiber shortening (corrected for heart rate)

LVEDD

Left ventricular end-diastolic diameter

WS

Wall stress

SVR

Systemic vascular resistance

MPI

Myocardial performance index

HR

Heart rate

Appendix: PDA Ligation/Hypotension Trial Investigators

Study Coordinating Center:

University of California San Francisco, San Francisco, CA

Ronald Clyman, MD

Andrea Wickremasinghe, MD

Michael Brook, MD

Neonatal Clinical Research Center nurses

Steroid Measurement Center:

University of Michigan, Ann Arbor, MI

Richard Auchus, MD

Susan Matthews, PhD

Robert Chomic, MS

Stephen C. Brown, PhD

Study Sites:

Loma Linda University, Loma Linda, CA (n=23)

T. Allen Merritt, MD, MHA

Tabitha Solomon, MD

Michael Kuhn, MD

Hospital for Sick Children, Toronto, Canada (n=16)

Patrick McNamara, MD

Amish Jain, MD

Afif El-Khuffash MD

Jenny Luc, RRT

University of Oklahoma, Oklahoma City, OK (n=6)

Shahab Noori, MD

Krishnamurthy Sekar, MD

Michael McCoy, MS, APRN

Karen Corff, MS, APRN

Children’s & Women’s Health Centre of British Columbia, Vancouver, Canada (n=6)

Pascal Lavoie, MD, PhD

Jennifer Claydon, MSc

Nadine Lusney, RN

Kristi Finlay, RN

BC Children's Hospital Cardiology service

University of Virginia, Charlottesville, VA (n=5)

Joshua T. Attridge, MD

Amy E. Blackman, RN

Jonathan R. Swanson, MD

Pediatric Cardiac Sonographers

Vanderbilt University, Nashville, TN (n=4)

Maria Gillam-Krakauer, MD

Jeff Reese, MD

Amy Law-Beller, RN

Indiana University, Indianapolis, IN (n=2)

Brenda Poindexter MD

Leslie Dawn Wilson, BSN, CCRC