Levosimendan, milrinone, and dobutamine in experimental acute pulmonary embolism.

Acute pulmonary embolism is a frequent condition in emergency medicine and potentially fatal. Cause of death is right ventricular failure due to increased right ventricular afterload from both pulmonary vascular obstruction and vasoconstriction. Inodilators are interesting drugs of choice as they may improve right ventricular function and lower its afterload. We aimed to investigate the cardiovascular effects of three clinically relevant inodilators: levosimendan, milrinone, and dobutamine in acute pulmonary embolism. We conducted a randomized, blinded, animal study using 18 female pigs. Animals received large autologous pulmonary embolism until doubling of baseline mean pulmonary arterial pressure and were randomized to increasing doses of each inodilator. Effects were evaluated with bi-ventricular pressure-volume loop recordings, right heart catheterization, and blood gas analyses. Induction of pulmonary embolism increased right ventricular afterload and pulmonary pressure (p < 0.05) causing right ventricular dysfunction. Levosimendan and milrinone showed beneficial hemodynamic profiles by lowering right ventricular pressures and volume (p < 0.001) and improved right ventricular function and cardiac output (p < 0.05) without increasing right ventricular mechanical work. Dobutamine increased right ventricular pressure and function (p < 0.01) but at a cost of increased mechanical work at the highest doses, showing an adverse hemodynamic profile. In a porcine model of acute pulmonary embolism, levosimendan and milrinone reduced right ventricular afterload and improved right ventricular function, whereas dobutamine at higher doses increased right ventricular afterload and right ventricular mechanical work. The study motivates clinical testing of inodilators in patients with acute pulmonary embolism and right ventricular dysfunction.

Introduction

Acute pulmonary embolism (PE) is a life-threatening condition that frequently needs emergency and intensive care. Acute PE increases right ventricular (RV) afterload causing RV failure and death. The increase in RV afterload is due to a combination of mechanical obstruction and vasoconstriction in the pulmonary circulation.

Accordingly, inodilators are an interesting class of drugs as their pharmacodynamic profile may improve RV function and decrease RV afterload. However, only pulmonary vasoconstriction, and not the mechanical obstruction, is susceptible to a vasodilating agent making it difficult to predict the efficacy of inodilators to reduce RV afterload in acute PE.

Clinically available inodilators include levosimendan, milrinone, and dobutamine, all of which have different hemodynamic profiles.[3-6] Inodilators are used in acute cardiac care and in chronic pulmonary hypertension (PH),[3,7] but only case reports or small case series exist on their effect in acute PE.[8,9] At present, dobutamine has received a Class IIa recommendation limited to high-risk PE patients, whereas neither levosimendan nor milrinone is mentioned in the European Society of Cardiology Pulmonary Embolism guidelines.

Previous animal studies have mostly investigated a single drug[10-15] and more research is warranted to investigate the circulatory supportive effects of inodilators in acute PE. We aimed to investigate the hemodynamic profiles of levosimendan, milrinone, and dobutamine in a porcine model of autologous acute intermediate-risk PE. We hypothesized that all inodilators would lower RV afterload and improve RV function in acute PE.

Methods

Design

After instrumentation and stabilization, we conducted a baseline measurement. Pulmonary emboli were introduced until doubling of mean pulmonary arterial pressure (mPAP). The PE measurement was conducted 30 min after the last emboli. Animals were then randomized by draw to four logarithmically increasing doses (D1–D4), 30 min of each dose, of either levosimendan, milrinone, or dobutamine (Fig. 1). Study was ended by euthanasia from a lethal dose of pentobarbital.

Fig. 1.

Study design. Animals were anesthetized and instrumented before baseline evaluation. Pulmonary emboli were administered until doubling of pulmonary pressure from baseline. Animals were randomized to either levosimendan, milrinone, or dobutamine. For all inodilators, four logarithmically increasing doses were administered for 30 min each.

Drugs

For all drugs, dosages below, within, and above clinical ranges were chosen with a logarithmic increase. Levosimendan was loaded with 2.4, 6, 24, and 60 µg/kg for 10 min followed by 20 min of continuous infusion of 0.02, 0.05, 0.2, and 0.5 µg/kg/min, respectively. Milrinone was loaded with 10, 30, 100, and 300 µg/kg loading doses for 10 min followed by concomitant infusion over 20 min of 0.1, 0.3, 1, and 3 µg/kg/min, respectively. Dobutamine was administered as continuous infusion of 1, 3, 10, and 30 µg/kg/min with no loading dose, but an equal amount of isotonic glucose solution was administered to exclude any volume effects.

Animals and ethics

We used female, Danish pigs of 58.1 ± 2.9 kg. Animals were handled in accordance with Danish legislation on animal welfare and care. Approval was obtained from the Danish Animal Research Inspectorate (no.: 2016-15-0201-00840). Animals were excluded if they at baseline had mPAP >25 mmHg, as clinical cutoff for PH. The study compiled with the ARRIVE guidelines. In accordance with the 3R recommendations on animal research, we previously conducted a thorough study on cardiopulmonary changes in our model acting as a reference group for several interventional protocols including the current study. The reference study showed no cardiovascular changes in the early hours following acute PE, i.e. in the time frame of the present study.

Anesthesia and instrumentation

The model is described in Supplementary Methods in detail and has been described previously. Of note, we inserted a 26F sheath (Dry-Seal, Gore Medical, USA) in the external jugular vein for administration of large emboli en bloc. Bi-ventricular pressure–volume (PV) loop data (emka Technologies, Paris, France) were recorded in LabChart through admittance control units (ADV 500, Transonic Scisense, London, Canada) and PowerLab 8/35 (ADInstruments, Oxford, UK). PV data were analyzed in LabChart with the observer blinded to the randomization group and time point.

PE model

Each PE was created from 30 mL of the animals’ own blood (size 20 × 1 cm) to cause central PE.[16,17] The PE model is described in details in Supplementary.

Measurements

We recorded mPAP, central venous pressure (CVP) by right heart catheterization, and mean arterial pressure (MAP) in the femoral artery. From the PV setup, we recorded RV and left ventricular (LV) end-systolic and end-diastolic pressures and volumes (ESP, EDP, ESV, and EDV, respectively) as well as ejection fraction (EF), stroke volume (SV), arterial elastance (Ea), cardiac output (CO), and first derivatives of pressure (dP/dt(max)). Load-independent data were recorded during a transient inferior vena cava occlusion (e.g. end-systolic elastance (Ees), preload recruitable stroke work, and pressure–volume area normalized to EDV (PVA/EDV), a measure of total mechanical work and surrogate for oxygen consumption). See Supplementary Methods for further details.

Statistics

We estimated a possible reduction in pulmonary vascular resistance (PVR) in our model from 300 to 200 dynes × s × cm−5 (SD 50) based on previous experience, a reduction also being a clinically relevant reduction. By comparing two means with α = 0.05 and β = 0.80, we calculated a sample size of n = 6 per group. The different pharmacokinetic properties did not allow equipotent administration and that is why the study was not powered for statistical comparison between groups. We used Shapiro–Wilks’ test and QQ-plot for normal distribution evaluation. If results are normally distributed, they are presented as mean ± SD, otherwise median (interquartile range). The three groups were compared at baseline using one-way ANOVA and Tukeys’ post-hoc multiple comparison test or Kruskal–Wallis test with Dunn’s correction as appropriate. The PE model was evaluated by paired t-test or Wilcoxon signed-rank-test as appropriate comparing baseline and PE measurements in each of the groups. Effects of the pharmacological interventions were calculated as the absolute difference from the PE time point and evaluated by repeated measurement one-way ANOVA with test-for-trend to include the order of measurements. Friedman’s test was used on non-normally distributed data. To control power in multiple testing, we used hierarchical testing from D4 and backwards, comparing measurements to the PE time point using paired t-test or Wilcoxon signed-rank-test where appropriate. A p-value < 0.05 was considered statistically significant. Prism 8.4.1 (GraphPad Software, LCC, CA) and Stata 15.1 (StataCorp, TX) were used for calculations.

Results

Eighteen animals were included and randomized 1:1:1 to the three inodilators. Animals were comparable at baseline confirming successful randomization (Table 1 and LV function in Supplementary Table 1).

Table 1

Cardiopulmonary physiology at baseline and after acute pulmonary embolism.

	Baseline				Pulmonary embolism
Variable	Levosimendan	Milrinone	Dobutamine	p	Levosimendan	Milrinone	Dobutamine
Hemodynamics
Heart rate, min–1	59 ± 11	64 ± 21	53 ± 11	0.46	69 ± 13	69 ± 14	65 ± 12
MAP, mmHg	77 ± 8	81 ± 6	71 ± 11	0.15	81 ± 6	80 ± 11	72 ± 14
CO, mL/min	5066 ± 871	7431 ± 3320	4608 ± 1095	0.07	5328 ± 944	6482 ± 1896	4218 ± 1390
Pulse oximetry, %	100 (99.8–100)	100	100	N/A	99 (97–100)	100 (99–100)	99 (96–100)
mPAP, mmHg	13.5 (12.9–15.8)	12.9 ± 1.7	13.8 ± 1.7	0.42	28.9 ± 6.3a	30.3 ± 5.8b	29.2 ± 3.8c
CVP, mmHg	4 ± 1	2 ± 2	5 ± 1e	0.025	5 ± 3	5 ± 3	6 ± 1
Blood lactate, mmol/L	0.83 ± 0.30	0.65 ± 0.16	0.80 ± 0.20	0.36	0.70 ± 0.23a	0.65 ± 0.19	0.97 ± 0.53
PVR, dynes × s × cm–5	69 ± 41	35 ± 31	93 ± 75	0.19	289 ± 129b	265 ± 128b	430 ± 194b
SVR, dynes × s × cm–5	1187 ± 267	979 ± 369	1210 ± 404	0.47	1154 ± 177	998 ± 343	1393 ± 644
PVR/SVR	0.06 ± 0.03	0.03 ± 0.02	0.07 ± 0.04	0.11	0.26 ± 0.12b	0.27 ± 0.10b	0.32 ± 0.10c
Ventilation
pH	7.49 ± 0.01	7.49 ± 0.02	7.47 ± 0.02	0.29	7.42 ± 0.03b	7.40 ± 0.04c	7.36 ± 0.04c
PaCO2, kPa	5.0 ± 0.2	5.3 ± 0.3	5.3 ± 0.3	0.20	6.2 ± 0.3c	6.7 ± 0.7c	7.1 ± 0.6c
PaO2, kPa	19.6 ± 1.0	19.9 ± 1.7	20.1 ± 1.6	0.79	13.5 ± 3.5b	13.3 ± 3.8b	11.5 ± 1.4d
PvO2, kPa	5.2 ± 0.9	6.2 ± 0.7	5.3 ± 0.8	0.10	5.6 ± 0.8	5.1 ± 0.7a	5.0 ± 0.4
Arterial saturation, %	99 (99–100)	100 (99–100)	100 (99–100)	0.82	96 ± 3a	96 ± 5 (p = 0.06)	94 ± 3a
Arterial O2 content, mL/dL	12.7 ± 0.5	13.3 ± 0.7	13.3 ± 0.7	0.24	13.1 ± 0.7	12.9 ± 0.7	13.2 ± 1.1
Physiological dead space	–18 (–48 to –14)	9 ± 44	–8 ± 28	0.15	91 ± 47a	167 ± 72d	163 ± 57c
EtCO2, kPa	5.3 ± 0.2	5.2 ± 0.2	5.3 ± 0.4	0.65	5.1 ± 0.4	4.6 ± 0.5b	4.9 ± 0.6a
Right ventricular function
RV ESP, mmHg	27.8 ± 2.9	26.4 ± 3.4	27.9 ± 3.5	0.69	46.0 ± 7.6b	49.3 ± 10.9b	49.5 ± 6.1d
RV ESV, mL	24 ± 15	27 ± 10	25 ± 13	0.95	40 ± 15b	47 ± 18a	51 ± 18c
RV EDV, mL	140 ± 26	146 ± 22	147 ± 32	0.90	153 ± 33a	154 ± 39	167 ± 27
RV EF, %	76 ± 13	75 ± 7	77 ± 7	0.92	58 ± 11b	55 ± 7b	52 ± 5d
RV Ea, mmHg/mL	0.32 ± 0.03	0.24 ± 0.07	0.35 ± 0.13	0.11	0.59 ± 0.03b	0.53 ± 0.22a	0.83 ± 0.35b
RV SW, mmHg × mL	1708 ± 243	1880 ± 365	1769 ± 395	0.68	1921 ± 608	1748 ± 671	2005 ± 423a
RV dP/dtmax, mmHg/s	403 ± 115	365 ± 61	359 ± 27	0.58	519 ± 133b	459 ± 83b	513 ± 111a
RV Ees, mmHg/mL	0.47 ± 0.21	0.35 ± 0.14	0.45 ± 0.12	0.40	0.63 ± 0.16a	0.48 ± 0.12	0.63 ± 0.23a
RV PRSW, mmHg	12.7 ± 3.8	15.0 ± 1.8	13.8 ± 2.0	0.36	15.3 ± 6.9	13.5 ± 6.0	11.7 ± 3.4
RV PVA/EDV	16.8 ± 2.3	19.2 ± 5.3	15.9 ± 1.6	0.27	26.3 ± 5.6b	26.2 ± 4.9a	21.3 ± 4.9a
RV Ees/Ea	1.51 ± 0.71	1.45 ± 0.51	1.36 ± 0.54	0.91	1.07 ± 0.51	0.96 ± 0.23	0.91 ± 0.68b

Notes: Hemodynamics at baseline with p-values from one-way ANOVA. Changes after acute pulmonary embolism induction compared to the baseline value of each of the three groups. n = 6 for all.

p < 0.05.

p < 0.01.

p < 0.001.

p < 0.0001 vs baseline. ep<0.05 vs milrinone group.

MAP: mean arterial pressure; CO: cardiac output; mPAP: mean pulmonary arterial pressure; CVP: central venous pressure; PVR: pulmonary vascular resistance; SVR: systemic vascular resistance; PaCO2: arterial partial pressure of CO2; PaO2: arterial pressure of O2; PvO2: venous partial pressure of O2; RV: right ventricular; ESP: end-systolic pressure; ESV: end-systolic volume; EDV: end-diastolic volume; EF: ejection fraction; Ea: arterial elastance; SW: stroke work; Ees: end-systolic elastance; PRSW: preload recruitable stroke work; PVA: pressure–volume area; EtCO2: end-tidal carbon dioxide.

Effects of PE

Animals received 5 ± 1 PE to reach the target of two times baseline mPAP. Hemodynamic changes are presented in Table 1. PE caused increases in RV afterload evident by an increase in mPAP (p < 0.05), PVR (p < 0.01), PVR to systemic vascular resistance (SVR), and RV Ea (p < 0.05). Preload increased as measured by a CVP increase (with p-values 0.05–0.06 in all groups). CO and MAP were unaltered. RV volume (p < 0.05) and ESP (p < 0.01) increased. RV systolic function was attenuated showing a decreased RV EF (p < 0.01) despite increases in measures of myocardial contractility RV Ees and dP/dt(max). RV PVA/EDV expressing total mechanical work increased (p < 0.05). Collectively, a non-significant RV ventriculo–arterial uncoupling with approximately one-third reduction in RV Ees/Ea was observed. On ventilatory function, we observed decreased PaO2 and saturation, and increased PaCO2 and physiological dead space (Table 1). Effects on RV diastolic function and LV function are presented in Supplementary Table 1. In general, acute PE did not significantly affect these.

Effects of levosimendan

Levosimendan increased CO (p < 0.0001), HR, and SV, and decreased CVP, SVR, and MAP. RV afterload (PVR, RV Ea, and mPAP) decreased. See Table 2 or Fig. 2. A bi-modal curve of PVR/SVR was noted (Fig. 2(d)). On RV function, levosimendan decreased pressures (p = 0.007) and volumes and increased EF (p = 0.049). RV contractility increased with improved RV coupling (p = 0.0058) without effects on mechanical work (Fig. 3 and Table 2). Effects on ventilatory function, RV diastolic function, and LV function are presented in Supplementary Tables 2 and 3 and Supplementary Figure 1. In general, levosimendan improved LV function and LV–arterial coupling even with decreased mechanical work but had limited effect on blood gasses.

Table 2.

Effects of inodilators on hemodynamics and right ventricular function.

Variables	Drug	D1	D2	D3	D4	p-Values
Hemodynamics
Heart rate, bpm	Levo	–3 ± 4	3 ± 11	15 ± 17	22 ± 11b	<0.0001
	Mil	2 ± 6	13 ± 18	31 ± 25a	39 ± 14c	<0.0001
	Dob	3 ± 7	15 ± 13a	66 ± 17c	98 ± 17d	<0.0001
Stroke volume, mL	Levo	0 ± 3	1 ± 7	9 ± 17	19 ± 31	0.0210
	Mil	–7 ± 6	0 ± 19	0 ± 30	–5 ± 21	0.8654
	Dob	9 ± 11	6 ± 15	–3 ± 15	–10 ± 13	0.0304
Mean arterial pressure, mmHg	Levo	–1 ± 2	–2 ± 1b	–3 ± 2a	–8 ± 6a	<0.0001
	Mil	–2 ± 10	–1 ± 11	–6 ± 14	–17 ± 15a	0.0003
	Dob	2 ± 4	8 ± 5	6 ± 6	–7 ± 8	0.0998
Systemic vascular resistance, dynes × s × cm–5	Levo	36 ± 64	–59 ± 172	–296 ± 231a	–470 ± 245b	<0.0001
	Mil	33 ± 146	–36 ± 101	–257 ± 182a	–466 ± 152c	<0.0001
	Dob	–188 ± 122	–289 ± 347	–627 ± 329b	–791 ± 378b	<0.0001
Central venous pressure, mmHg	Levo	–1 ± 2	–2 ± 1b	–2 ± 1b	–3 ± 1b	<0.0001
	Mil	–1 ± 1a	–1 ± 1a	–2 ± 1b	–4 ± 1c	<0.0001
	Dob	–1 ± 1	–1 ± 1a	–3 ± 3a	–4 ± 3a	0.0412
Blood lactate, mmol/L	Levo	–0.05 ± 0.05	–0.12 ± 0.10	–0.12 ± 0.15	0.00 ± 0.17	0.5677
	Mil	0.07 ± 0.37	0.08 ± 0.61	0.07 ± 0.67	0.05 ± 0.68	0.7941
	Dob	–0.22 ± 0.26	–0.40 ± 0.39	–0.45 ± 0.50	–0.30 ± 0.51	0.0081
Urine production, mL/h	Levo	60 (53–143)	79 (55–333)	159 (63–520)	120 (62–255)	0.2928
	Mil	101 ± 60	104 ± 38	86 ± 42	87 ± 27	0.4720
	Dob	50 (19–113)	65 (45–139)	102 ± 118	98 ± 95	0.2469
RV function
RV ESV, mL	Levo	–2 ± 5	–7 ± 9	–12 ± 7b	–16 ± 5c	<0.0001
	Mil	–5 ± 5	–9 ± 7a	–14 ± 11a	–20 ± 9b	<0.0001
	Dob	–8 ± 8	–23 ± 8c	–35 ± 11c	–38 ± 14b	<0.0001
RV EDV, mL	Levo	–14 ± 15	–33 ± 28a	–50 ± 30b	–60 ± 33b	<0.0001
	Mil	–20 ± 22	–33 ± 32	–49 ± 36a	–60 ± 40a	<0.0001
	Dob	–38 ± 32a	–73 ± 44b	–113 ± 30c	–132 ± 36c	<0.0001
RV SW, mmHg × mL	Levo	–152 ± 310	–342 ± 518	–516 ± 866	–486 ± 1012	0.0318
	Mil	–27 ± 447	–101 ± 548	–202 ± 452	–627 ± 386a	0.0006
	Dob	–83 ± 652	–135 ± 572	–480 ± 557	–776 ± 650a	0.0107
RV Ees	Levo	0.02 ± 0.15	0.01 ± 0.14	0.10 ± 0.12	1.03 ± 1.82	0.0582
	Mil	0.07 ± 0.10	0.18 ± 0.16	0.22 ± 0.25	0.43 ± 0.29a	<0.0001
	Dob	0.10 ± 0.17	0.73 ± 0.48	1.41 ± 1.01	1.77 ± 2.13	0.0010
RV Ea, mmHg/mL	Levo	–0.04 ± 0.05	–0.06 ± 0.06	–0.11 ± 0.08	–0.13 ± 0.14	0.0006
	Mil	–0.01 ± 0.02	–0.03 ± 0.09	–0.08 ± 0.12	–0.09 ± 0.09	0.0124
	Dob	–0.13 ± 0.13	–0.08 ± 0.17	0.34 ± 0.34	0.71 ± 0.54a	<0.0001
RV dP/dt max, mmHg/s	Levo	–29 ± 45	–18 ± 60	24 ± 40	89 ± 133	0.0233
	Mil	–9 ± 42	6 ± 55	34 ± 80	91 ± 124	0.0067
	Dob	73 ± 101	318 ± 194	1096 ± 305	1759 ± 416	<0.0001
RV PRSW, mmHg	Levo	1.6 ± 3.6	3.3±5.5	1.8±5.6	6.5±4.6a	0.0250
	Mil	–0.8 ± 4.5	0.5±4.5	0.8±2.8	3.6±6.1	0.0498
	Dob	3.9 ± 6.2	8.7±4.9b	16.9±11.7a	32.3±17.0b	<0.0001

Notes: Absolute changes from PE time point to the four doses of each inodilator. p-Values represent the one-way ANOVA test-for-trend. n = 6 for all.

p < 0.05.

p < 0.01.

p < 0.001.

p < 0.0001 vs PE.

D1–D4: dose 1–4; RV: right ventricular; ESV: end-systolic volume; EDV: end-diastolic volume; SW: stroke work; Ees: end-systolic elastance; Ea: arterial elastance; PRSW: preload recruitable stroke work.

Fig. 2.

Cardiopulmonary effects of levosimendan, milrinone, and dobutamine in acute pulmonary embolism. Effects of inodilators are shown as absolute changes from PE time point. Different effects were noted on pulmonary pressure (a), despite each drug decreased calculated pulmonary vascular resistance (b), and increased cardiac output (c). Note the U-shaped curve of pulmonary to systemic resistance (d). Values are mean ± SD. n = 6 for each mean. Each p-value describes the overall test-for-trend. *p < 0.05, **p < 0.01, *** < 0.001 vs. PE.

PE: pulmonary embolism.

Fig. 3.

Effects of inodilators on right ventricular function. Effects of inodilators are shown as absolute changes from PE time point. Levosimendan and milrinone consistently decreased end-systolic pressure (a) and increased ejection fraction (b) with no effects on pressure–volume area relative to end-diastolic volume, i.e. total mechanical work (d). On the contrary, dobutamine increased the end-systolic pressure and increased mechanical work as surrogate for oxygen consumption. All increased RV ventriculo–arterial coupling (c). Values are mean ± SD. n = 6 for each mean. Each p-value describes the overall test-for-trend. *p < 0.05, **p < 0.01, *** < 0.001 vs. PE.

PE: pulmonary embolism.

Effects of milrinone

Milrinone increased CO (p < 0.0001), mostly through increased HR (p < 0.0001), and lowered CVP, SVR, and MAP. Decreases in RV afterload estimates (mPAP, PVR, and RV Ea) were observed (Table 2 and Fig. 1). PVR/SVR showed a bi-modal curve (Fig. 2(d)). Milrinone decreased RV pressure and volume, increased contractility (p < 0.0001), and improved RV coupling (p < 0.0001) (Table 2 and Fig. 3). Mechanical work was unaffected. LV systolic and diastolic function and LV–arterial coupling were improved, whereas only minor effects on ventilatory function were observed (Supplementary Tables 2 and 3 and Supplementary Figure 1).

Effects of dobutamine

Dobutamine caused pronounced tachycardia (p < 0.0001), as the main contributor to increased CO. Dobutamine lowered SVR but MAP was unaffected. Dobutamine decreased blood lactate levels. Measures of RV afterload showed divergent results as PVR decreased (p < 0.0001), mPAP was unaltered but RV Ea increased (p < 0.0001) (Fig. 2 and Table 2). Dobutamine did not improve RV function in a stepwise manner; instead increased RV systolic pressure and reductions in SV and RV EF were observed, especially at higher doses (Fig. 3). RV contractility was increased by dobutamine also improving RV–PA coupling (p = 0.0042) despite increased afterload. The increased contractility was at a cost of increased mechanical work (p < 0.0001). See Table 2 and Fig. 3. Similarly, the LV pressure and contractility increased without improvements in LV coupling and at a cost of increased mechanical work (Supplementary Figure 1 and Supplementary Table 2). Dobutamine increased O2 content and O2 delivery (Supplementary Table 3).

Discussion

We induced autologous, acute PE in pigs causing increased RV afterload. RV dysfunction was evident by RV dilatation, decreased RV EF, and RV–PA uncoupling. As systemic blood pressure was preserved, the model mimics intermediate-risk PE in clinical risk stratification. Levosimendan and milrinone reduced RV afterload and improved RV function, whereas dobutamine, especially at higher doses, increased CO but at a cost of increased RV afterload and mechanical work.

These effects are illustrated in Fig. 4 of representative RV PV loops from baseline and after acute PE where both volume and pressure increase. Furthermore, loops from animals treated with each of the drugs are shown; all drugs cause a leftwards shift of the loops corresponding to decreased RV volume, but only dobutamine increases RV pressures at the same time, whereas the RV PV loops in the levosimendan and milrinone groups are shifted left- and downwards.

Fig. 4.

Right ventricular pressure–volume loops. Representative right ventricular pressure–volume loops from baseline and after acute pulmonary embolism (top). Note the difference in shape where the normal right ventricle at baseline has a more triangular shape with pressure decrease during systolic ejection contrary to the end-systolic notch present after pulmonary embolism reflecting increased afterload. Levosimendan and milrinone cause left- and downward shift in pressure–volume loops, whereas dobutamine causes leftwards shift and increased pressure. For simplicity of the figure, only two doses are shown.

Few animal studies have evaluated all three inodilators together in different models, mostly favoring levosimendan.[19-23] Only one clinical study has compared the inodilators in 60 patients with postoperative low CO syndrome, where milrinone performed less well. The present study is the first to include all three drugs in acute PE and one of the few to include load-independent data from PV loop recordings.

Levosimendan in acute PE

Levosimendan, especially at low doses, had a favorable hemodynamic profile with a reduction in RV afterload and modest improvement in RV function without increasing total mechanical work as predicted from its known pharmacodynamic properties.[3,5] Our observations confirm previous findings in acute PE. We have previously shown in a chronic animal model that the improvement in RV function seen with levosimendan mainly depends on its ability to induce pulmonary vasodilatation.

The observed decrease in PVR can be explained by the increase in CO and interpretation of changes in PVR must be done with caution when CO changes.[15,25] However, when excluding CO from the analysis by normalizing PVR to SVR as seen in Fig. 2(d), we noticed a bi-modal curve with the largest reductions at low levosimendan concentrations. These findings indicate that levosimendan is a more selective pulmonary vasodilator at low doses in acute PE.

Urine production was numerically larger in the levosimendan group but did not show dose–response relationship. Levosimendan is known to have renal-protective effects in a range of surgical and critical illness situations, including low-output stages. Similarly, CO may decrease in acute PE.

Milrinone in acute PE

Milrinone showed a favorable hemodynamic profile with decreased RV afterload and improved RV function without affecting mechanical work. A bi-modal PVR/SVR curve again suggested milrinone to be a more selective pulmonary vasodilator at low doses. Milrinone did, however, lower SVR and MAP which can be a potential clinical concern in patients with acute PE and low-to-normal blood pressure. These findings are comparable to other’s findings on acute[11,12,27-29] and chronic models.[22,30,31] As for levosimendan, RV afterload-reducing effects of milrinone seem to depend on pre-constricted conditions as mPAP and PVR are unaffected in normal pulmonary vasculature.[12,19] The present study shows moderate dosage of milrinone to be safe in intermediate-risk PE with the numerical largest effect on mPAP and low effect on MAP. Milrinone, compared with levosimendan, showed a more prominent chronotropic effect.

Dobutamine in acute PE

Dobutamine had the most pronounced effects on CO, RV contractility, and especially heart rate. Contrary to the other inodilators, dobutamine increased RV pressure, af and mechanical work (i.e. oxygen consumption). Fig. 2(d) indicates dobutamine to be the least selective pulmonary vasodilator as previously suggested.[14,15,27]

Unwanted effects from high dobutamine doses have been shown previously. We noticed a flipped U-curve relationship, and not a direct dose–response relationship, on RV EF and SV. This can be explained by tachycardia that reduced ventricular filling time to an extent that compromised EF and SV. Similarly, the RV afterload increase outweighed the increase in RV contractility failing to improve RV–PA coupling at higher doses. Similar patterns were observed for LV.

We confirmed dobutamine’s ability to increase CO but inability to reduce mPAP in acute PE.[13-15,27] The PVR reduction should be interpreted by the CO increase rather than afterload reduction as the RV Ea actually increased. Our observed hemodynamic profile is in concordance with models of normal pulmonary vasculature or acute or chronic RV failure.[19-21,25,31-33] The unwanted increase in mechanical work and oxygen consumption has been shown previously[19,22,32] and was also expected from the pharmacodynamic properties.[5,6] Oxygen consumption only increased in the myocardium. Dobutamine did also increase oxygen delivery and concomitant to the washout effect of increased CO, the blood lactate decreased. The lactate levels in the dobutamine group were, however, higher following PE increasing the magnitude of the response.

The present data show positive effects on RV contractility, RV–PA coupling, and function at low doses before chronotropic side effects become evident, which is a known dosage limitation of dobutamine. At low doses, dobutamine tended to increase MAP supporting its recommendation in high-risk PE with systemic hypotension. However, the present study showed decreasing SVR and increasing heart rate and RV afterload which might be unwanted side effects in high-risk PE.

Ventilatory function

In general, we observed either no or limited effects on blood gasses. This is in concordance with several other studies of the inodilators[10-12,14,22,27,34,35] and suggests against using blood gasses for evaluation of inodilators in acute PE.

Clinical considerations

Patients with acute PE might present with hemodynamic instability or in stable conditions with the risk of rapid deterioration. Vasopressors like norepinephrine or epinephrine are commonly used in these situations for systemic support. However, the cause of hemodynamic instability is RV failure, and vasopressors might induce pulmonary vasoconstriction with further RV strain as an unwanted side effect. Theoretically, inodilators could avoid such, and based on the present study, levosimendan (or milrinone) would be the drug of choice as both showed a safe hemodynamic profile in low doses without increasing RV strain. Inodilators should be saved for PE patients with manifest RV failure as those drugs might have harmful side effects, too, and should always be administered with a vasopressor support available. Based on the present study, further research should investigate effects of levosimendan or milrinone and a vasopressor in combination and clinical testing should be considered.

Limitations

The present study has some limitations to be addressed. As the study was performed on pigs in anesthesia and on mechanical ventilation, the model differs from clinical settings of acute PE and translation must be done with caution. However, our model of autologous, acute PE with RV dysfunction but preserved systemic blood pressure mimics intermediate-risk PE in clinical risk stratification and is therefore relevant in translational research. A control group was not part of the present study in accordance with animal research ethics. We have previously shown in detail the temporal development of hemodynamics and RV function in this model without interventions. In that study, the RV was dysfunctional throughout the time window of the present study, i.e. changes can be attributed the inodilators.

An effect of inodilators on clot lysis in this model cannot be excluded. Levosimendan may have pro-fibrinolytic effects, whereas dobutamine does not seem to have fibrinolytic effects. But fibrinolytic effects of inodilators have not been investigated in acute PE. Furthermore, pharmacokinetics and pharmacodynamics of the inodilators might have interspecies differences. Our design, however, included clinically relevant doses[4,5] showing relevant hemodynamic effects, and chosen doses correspond to other animal studies.[10,19,21,22,31] Lastly, measurements were performed after 30 min of infusion. Levosimendan has active metabolites with very long half-life and therefore the effects of levosimendan may be even more pronounced at another time window. However, we have previously shown effects on RV function after 10 min of levosimendan infusion and aimed for acute effects of the drugs in perspective of the immediate therapeutic demand in patients with acute PE. Future investigations should elucidate prolonged effects of inodilators in acute PE.

In conclusion, we investigated the cardiovascular effects of three inodilators: levosimendan, milrinone, and dobutamine in a porcine model of acute PE mimicking intermediate-risk PE. Levosimendan and milrinone showed similar and beneficial hemodynamic profiles with reduced RV afterload and improved RV function without increasing RV mechanical work. Dobutamine increased CO, but at the cost of an increase in RV afterload and RV mechanical work especially at higher doses. The study motivates clinical testing of inodilators in patients with acute PE and RV dysfunction.

Click here for additional data file.

Supplemental material, sj-pdf-1-pul-10.1177_20458940211022977 for Levosimendan, milrinone, and dobutamine in experimental acute pulmonary embolism by Mads D. Lyhne, Simone J. Dragsbaek, Jacob V. Hansen, Jacob G. Schultz, Asger Andersen and Jens Erik Nielsen-Kudsk in Pulmonary Circulation

Click here for additional data file.

Supplemental material, sj-pdf-2-pul-10.1177_20458940211022977 for Levosimendan, milrinone, and dobutamine in experimental acute pulmonary embolism by Mads D. Lyhne, Simone J. Dragsbaek, Jacob V. Hansen, Jacob G. Schultz, Asger Andersen and Jens Erik Nielsen-Kudsk in Pulmonary Circulation

Click here for additional data file.

Supplemental material, sj-pdf-3-pul-10.1177_20458940211022977 for Levosimendan, milrinone, and dobutamine in experimental acute pulmonary embolism by Mads D. Lyhne, Simone J. Dragsbaek, Jacob V. Hansen, Jacob G. Schultz, Asger Andersen and Jens Erik Nielsen-Kudsk in Pulmonary Circulation