OUP user menu

Effect of nitric oxide synthase inhibition on haemodynamics and outcome of patients with persistent cardiogenic shock complicating acute myocardial infarction: a phase II dose-ranging study

Vladimir Džavík, Gad Cotter, Harmony R. Reynolds, John H. Alexander, Krishnan Ramanathan, Amanda L. Stebbins, David Hathaway, Michael E. Farkouh, E. Magnus Ohman, David A. Baran, Roland Prondzinsky, Julio A. Panza, Warren J. Cantor, Zvi Vered, Christopher E. Buller, Neal S. Kleiman, John G. Webb, David R. Holmes, Joseph E. Parrillo, Stanley L. Hazen, Steven S. Gross, Robert A. Harrington, Judith S. Hochman
DOI: http://dx.doi.org/10.1093/eurheartj/ehm075 1109-1116 First published online: 25 April 2007


Aims Previous studies suggested haemodynamic benefits and, possibly, mortality reduction with the use of nitric oxide synthase (NOS) inhibition in patients with acute myocardial infarction (AMI) complicated by cardiogenic shock (CS). We assessed preliminary efficacy and safety of four doses of l-n-monomethyl-arginine (l-NMMA), a non-selective NOS inhibitor, in patients with AMI complicated by CS despite an open infarct-related artery.

Methods and results Patients (n = 79) were randomly assigned to a bolus and 5 h infusion of placebo or 0.15, 0.5, 1.0, or 1.5 mg/kg of l-NMMA. The primary outcome measure was absolute change in mean arterial pressure (MAP) at 2 h. Fifteen minutes after study drug initiation, mean change in MAP was −4.0 mmHg in the placebo group and 5.8 (P = 0.02), 4.8 (P = 0.02), 5.1 (P = 0.07), and 11.6 (P < 0.001) mmHg in the four l-NMMA groups, respectively (all vs. placebo). Mean change in MAP at 2 h was −0.4, 4.4, 1.8, −4.1, and 6.8 mmHg in the placebo and four l-NMMA groups, respectively (all P = NS).

Conclusion l-NMMA resulted in modest increases in MAP at 15 min compared with placebo but there were no differences at 2 h.

  • Acute myocardial infarction
  • Cardiogenic shock
  • Inflammation
  • Nitric oxide synthase
  • l-n-monomethyl-arginine (l-NMMA)


Cardiogenic shock (CS) remains the leading cause of death in patients hospitalized for acute myocardial infarction (AMI).1 Despite reductions in the intermediate and long-term mortality by early revascularization demonstrated by the SHOCK Trial,24 long-term survival remains poor, with mortality in excess of 50%.1 An overlap between CS and the systemic inflammatory response syndrome characterized by cytokine release and the expression of inducible nitric oxide synthase (iNOS) leading to high levels of nitric oxide has recently been recognized.5,6 In the healthy state, NO is continuously produced from l-arginine at low concentrations by a calcium-dependent NOS control of normal vascular tone at the level of the endothelium.7 At higher levels, NO causes inappropriate systemic vasodilatation as well as systemic and coronary hypoperfusion and may contribute directly to myocardial depression.5,8,9 NG-monomethyl-l-arginine (l-NMMA) is a reversible, non-selective NOS inhibitor that binds at the active catalytic site in competition with the normal substrate, l-arginine.7 Two small, single-centre studies suggested that non-selective NOS inhibition may improve selected clinical parameters, including blood pressure (BP), as well as survival of patients with CS.10,11

On the basis of the promising experimental and single-centre study findings, we conducted the SHould we inhibit nitric Oxide synthase in Cardiogenic shocK 2 (SHOCK-2) phase II dose-ranging study, which tested four doses of l-NMMA against placebo in patients with persistent CS complicating AMI. We assessed the preliminary safety and efficacy, pharmacokinetics, and biological activity of l-NMMA, as well as feasibility of a randomized trial of a pharmacological agent in the setting of AMI complicated by persistent CS.


Patient population

Patients with AMI complicated by persistent CS despite an open infarct-related artery (IRA) were considered for the study. Eligible patients had ECG evidence of ≥2 mm ST-segment elevation in at least two contiguous leads, left bundle branch block, or ≥2 mm ST-segment depression in at least two contiguous leads with elevated creatine kinase (CK), CK-MB, or troponin T or I above institutional upper limit of normal. Patients were eligible for randomization between 1 and 12 h after demonstration of an open IRA, defined as ≤70% stenosis with any Thrombolysis In Myocardial Infarction (TIMI) antegrade flow, recognizing that no reflow may be more likely to occur in this population. Patency could be achieved by percutaneous coronary intervention (PCI), thrombolytic therapy, or spontaneous reperfusion. Intra-aortic balloon counterpulsation (IABP) was strongly recommended. A pulmonary artery catheter was required.

CS criteria included clinical evidence of hypoperfusion with systolic BP < 100 mmHg despite at least 7 mcg/kg/min of dopamine or equivalent dose of norepinephrine in addition to low cardiac index (CI) (<2.2 L/min/m2 if measured off IABP, or <2.5 L/min/m2 on IABP) and pulmonary capillary wedge pressure (PCWP) ≥15 mmHg. These criteria were intended to select a population with expected mortality of at least 50%.3,12

Patients were excluded for suspected or documented infection, adult respiratory distress syndrome, primary pulmonary hypertension, proximal right coronary artery culprit (unless RV function was normal) or severe RV dysfunction from any cause (visualized on echocardiogram), other causes of shock (e.g. sepsis, hypovolaemia, haemorrhage, anaphylaxis), or CS due to causes other than left ventricular (LV) pump failure (severe mitral regurgitation, rupture of ventricular septum or ventricular free wall, brady- or tachyarrhythmia, or aortic dissection). Additional exclusion criteria included anticipated need for urgent surgical revascularization or LV assist device prior to 30 days post-randomization, serum creatinine >2.5 mg/dL (>221 µmol/L), anoxic brain damage, pre-terminal profound shock as evidenced by metabolic acidosis (pH < 7.1) on at least two measures over ≥30 min despite mechanical ventilation and circulatory support measures (e.g. IABP), irreversible multisystem failure (e.g. shock liver, lungs, kidneys), age < 18 years, or pre-existing illness with life expectancy < 6 months.

Institutional Review Board approval was obtained at all sites. Each subject or the subject's legally authorized representative gave written informed consent prior to participation.

Study design and procedures

SHOCK-2 was a phase II, randomized, placebo-controlled, dose-ranging study to assess preliminary efficacy and safety of four doses of l-NMMA compared with placebo. The primary outcome measure was change in MAP at 2 h after initiation of study drug. Secondary endpoints included change in MAP at 15 min and change in other haemodynamic variables at 15 min and 2 h. Additional endpoints included overall 30 day mortality and 30 day mortality for the pre-specified subgroups of patients <75 and ≥75 years of age. The study was designed to provide 80% power to detect a 20% increase in MAP and was not powered to detect differences in mortality.

Patients were randomly assigned by centralized interactive voice response system to one of five treatments: l-NMMA 0.15 mg/kg IV bolus and 0.15 mg/kg/h infusion × 5 h; 0.5 mg/kg IV bolus and 0.5 mg/kg/h infusion × 5 h; 1.0 mg/kg IV bolus and 1.0 mg/kg/h infusion × 5 h; 1.5 mg/kg IV bolus and 1.5 mg/kg/h infusion × 5 h; or placebo (0.9% normal saline), IV bolus, and 5 h infusion. Randomization was blocked and was stratified by site, with a projected randomization ratio of 1.33:1 for placebo vs. each active drug group. A screening log was maintained.

Baseline haemodynamics were recorded just prior to injection of study drug bolus. These and serial values included systolic and diastolic BP, MAP, PCWP, CI, cardiac power (CP), systemic vascular resistance (SVR), and pulmonary vascular resistance (PVR). BP measurements were taken from the in-dwelling arterial line or IABP visual display and by cuff after the removal of arterial line and IABP. All haemodynamic values were recorded on support measures, including vasopressors and inotropes, and with IABP paused, if clinically feasible. Down-titration of vasopressors was recommended for severe hypertension (systolic BP > 180 mmHg) or ischaemic events only, but clinicians also frequently down-titrated when hypotension was corrected.

The Duke Clinical Research Institute provided surveillance of safety and clinical events. An independent data and safety monitoring board reviewed study conduct on an ongoing basis. This study was conducted in accordance with the principles of Good Clinical Practice, including the World Medical Association's Declaration of Helsinki.

Central laboratory services were provided by ICON clinical laboratories in North America and Europe, and by American Medical Laboratories in Israel. Plasma concentrations of l-NMMA were assayed using a sensitive and specific liquid chromatography/mass spectrometry assay at The Cleveland Clinic Foundation. Samples were collected at baseline, 1, 5, 6, 12, and 24 h after study drug administration and were stored at −70°C prior to analysis.

Statistical analysis

All randomized subjects received assigned double-blinded study medication and are included in these analyses according to the assigned treatment. Dichotomous variables are presented as frequencies and continuous variables as means with standard deviation and, in the case of concomitant drug doses, time intervals and change in MAP at 15 min and 2 h, as medians with the 25th and 75th percentiles. On the basis of prior studies, standard deviations of the change in MAP of 10 mmHg for the placebo group and 20 mmHg for the active treatment groups were utilized for the sample size calculations.11

MAP measurements off IABP took precedence over measurements on IABP when both were available for the same time point; however, for paired data, consistency with regard to IABP status took precedence. Most paired measurements were performed with the IABP on.

The Dunnett test was the pre-specified test for statistical significance.13 Mean difference in each treatment arm was tested separately against placebo using a standard t-test at an α-level of 0.0146. For the pooled four active treatment groups, α-level was set at 0.05. A P-value of <0.0146 for any of the treatment arms was considered significant. A post hoc Jonckheere–Terpstra test was also performed on the primary outcome measure to account for the ordinal nature of the dosing. The four treatment arms were pooled and tested against placebo using a Fisher's exact test with mid-P-values. The quadratic trend test was also used to identify any linear or quadratic relationship across doses, assuming they were equally spaced. The Wilcoxon Rank Sum test was performed to detect if there is a significant difference in the median change in MAP at 15 min and 2 h, comparing each treatment arm to placebo. All dichotomous variables were analysed in this fashion. All statistical tests were two-sided.


Screening log data

Of 432 screened patients with suspected CS, 22 died before they could be enrolled, 52 did not meet haemodynamic inclusion criteria, 54 did not meet criteria for MI or IRA patency, and 116 had at least one exclusion criterion. The most frequent exclusion criteria met were renal dysfunction (n = 30), RV dysfunction (n = 22), shock duration>24 h (n = 14), and suspected infection (n = 13). A second exclusion criterion was present in half of patients with prolonged duration of shock. Shock resolved within 1 h of IRA opening in 81 patients; 10 patients were not identified within the 12 h window after confirmation of IRA patency. Of the 97 remaining patients who were eligible, 15 declined participation and three were treated with LV assist device in the early phase.

Baseline clinical characteristics

A total of 79 patients were enrolled at 22 centres in the USA (32), Canada (26), Germany (8), Israel (6), Austria (5), and Denmark (2) over 9 months. Treatment groups were similar with regard to age, sex, history of diabetes mellitus (DM), hypertension, and other comorbidities (Table 1). SHOCK-2 patients comprise an older population (age 69 ± 11 years) with a high proportion of DM, a finding consistent with prior studies of CS. With the exception of PCWP, there were no significant differences among groups in any baseline characteristic. Of note, there was variability between groups in baseline characteristics, which although not statistically significant, may be clinically relevant. This is not unusual for a dose-ranging study of this size. Median study intervals were similar across all groups (Table 2). PCI was performed prior to study drug administration in 90% of patients assigned to placebo and in 93, 93, 100, and 64% of patients assigned to 0.15, 0.5, 1.0, and 1.5 mg/kg of l-NMMA, respectively (P = 0.039). Median doses of vasopressors and inotropes at baseline just prior to study drug administration (Table 3) were lowest in the placebo group. Except for one patient in the 1.0 mg/kg group, all patients in SHOCK-2 received IABP support.

View this table:
Table 1

Baseline characteristics

Placebo (n = 20)l-NNMA 0.15 mg/kg (n = 15)l-NMMA 0.50 mg/kg (n = 15)l-NMMA 1.0 mg/kg (n = 15)l-NMMA 1.5 mg/kg (n = 14)
Age69.4 ± 11.168.4 ± 9.169.8 ± 10.966.8 ± 14.970.7 ± 8.1
Male sex (%)7067736064
Diabetes mellitus (%)4553534721
Hypertension (%)6567736071
Body surface area (m2)1.9 ± 0.31.9 ± 0.22.0 ± 0.31.9 ± 0.31.9 ± 0.2
Prior myocardial infarction (MI) (%)3027333314
Prior coronary artery bypass grafting (%)002070
Prior cardiovascular disease (%)1020700
Prior congestive heart failure (%)10132777
Chronic obstructive pulmonary disease (%)2027777
Shock onset [systolic blood pressure (BP)a]67.1 ± 11.873.1 ± 19.473.9 ± 12.865.5 ± 21.575.7 ± 6.1
Shock onset (diastolic BPa)42.2 ± 13.942.7 ± 13.747.2 ± 9.746.1 ± 15.345.6 ± 11.7
Qualifying mean arterial pressure (MAP)a63.8 ± 10.466.4 ± 13.768.4 ± 10.169.1 ± 8.667.7 ± 9.2
Baseline MAPa70.1 ± 13.466.3 ± 13.968.8 ± 9.373.0 ± 10.372.0 ± 8.8
Qualifying heart rate (HR)a91.3 ± 33.791.3 ± 33.289.1 ± 27.394.4 ± 30.987.0 ± 35.1
Qualifying pulmonary capillary wedge pressure (PCWP)a18.8 ± 2.923.1 ± 6.224.0 ± 6.720.9 ± 4.921.9 ± 6.1
Baseline systemic vascular resistance (SVR)a1503 ± 6451346 ± 5991418 ± 4611619 ± 5871642 ± 587
Baseline pulmonary vascular resistance (PVR)a220.1 ± 155.1123.5 ± 70.4190.4 ± 138.9151.5 ± 74.5330.2 ± 275.7
Qualifying cardiac index (CI)a1.6 ± 0.51.8 ± 0.61.7 ± 0.41.7 ± 0.51.7 ± 0.4
Left ventricular ejection fraction (LVEF) (%; n = 12, 7, 8, 9, 9)27.8 ± 9.025.6 ± 12.020.6 ± 7.629.2 ± 11.628.9 ± 9.9
Anterior MI (%)47.466.773.384.691.7
Baseline glucose (mg/dL)224.8 ± 120.3202.9 ± 118.2227.0 ± 106.8263.0 ± 198.9227.5 ± 161.3
  • All continuous variables are reported as mean ± standard deviation values.

  • aAll haemodynamic variables are measured on vasopressors, with some on and some off intra-aortic balloon counterpulsation (IABP). Baseline MAP measurements taken just prior to study drug administration off IABP support were available in eight patients in the placebo group, and five, six, eight, and 12 patients in the l-NMMA 0.15, 1.5, 1.0, and 0.50 mg/kg groups, respectively.

View this table:
Table 2

Study intervals

Placebo (n = 20)l-NMMA 0.15 mg/kg (n = 15)l-NMMA 0.50 mg/kg (n = 15)l-NMMA 1.0 mg/kg (n = 15)l-NMMA 1.5 mg/kg (n = 14)
Qualifying MI to shock onset (h)4.7 (1.7, 20.6)4.6 (1.7, 16.9)4.5 (1.0, 10.8)2.3 (1.7, 6.5)3.1 (0.9, 11.7)
Shock onset to randomization (h)9.1 (6.0, 11.7)8.5 (5.0, 14.5)6.1 (5.1, 13.0)9.2 (6.5, 12.9)7.9 (4.9, 15.6)
Open artery to randomization (h)4.0 (2.1, 8.6)4.1 (2.3, 6.6)3.9 (2.3, 8.4)5.9 (2.5, 8.1)4.7 (3.1, 5.9)
Patients receiving study drug (%)100100100100100
Randomization to study drug (h)0.9 (0.1, 2.2)0.9 (0.6, 2.8)1.0 (0.5, 2.8)0.8 (0.3, 2.4)1.2 (0.4, 1.9)
Duration of study drug infusion (h)5.0 (4.5, 8.0)5.0 (4.9, 11.5)5.0 (4.9, 5.1)5.0 (4.8, 7.7)5.0 (4.9, 6.0)
  • All times in hours, reported as median with 25th and 75th percentile, except for the duration of study drug infusion and randomization to study drug initiation, which is median, minimum, and maximum.

View this table:
Table 3

Median and 25th and 75th percentile doses of inotropes and vasopressors at baseline

Placebol-NMMA 0.15 mg/kgl-NMMA 0.5 mg/kgl-NMMA 1.0 mg/kgl-NMMA 1.5 mg/kg
Dopamine (µg/kg/min)1.50 (0.00, 7.25)6.76 (0.00, 12.00)7.00 (2.00, 9.00)7.00 (0.00, 10.00)3.50 (0.00, 8.00)
Dobutamine (µg/kg/min)0.00 (0.00, 4.02)0.00 (0.00, 2.40)0.00 (0.00, 3.00)0.00 (0.00, 5.00)0.00 (0.00, 2.50)
Norepinephrine (µg/kg/min)0.06 0.01, 0.13)0.02 (0.00, 0.12)0.00 (0.00, 0.01)0.01 (0.00, 0.18)0.02 (0.00, 0.12)
  • For epinephrine, vasopressin, milrinone, and phenylephrine, median and 25th and 75th percentile values were 0.

Pharmacokinetic and pharmacodynamic data

Following study drug administration, plasma drug levels were proportional to dose (Figure 1A). Concentrations of l-NMMA were similar at 1 and 5 h after initiation of study drug bolus and infusion, indicating that the selected loading dose was appropriately matched to the maintenance infusion for rapid attainment of l-NMMA steady state (Figure 1B). The steady-state concentration was linearly related to infusion rate (Figure 2). The median elimination half-life across the population (available for 49 subjects) was 0.64 h [IQR 0.52–0.91 h]. Of these measurements, 24% of the half-lives were <0.5 h; 41% were between 0.50 and 0.75 h; 12% were between 0.75 and 1.00 h; and 22% were >1 h.

Figure 1

(A) l-NMMA concentration–time profiles for the four treatment groups and placebo group. (B) Relationship of l-NMMA concentration at 1 and 5 h.

Figure 2

Steady-state l-NMMA concentration and infusion rate.

Primary outcome measure: change in mean arterial pressure at 2 h

Absolute change in MAP at 2 h was not significantly different in any l-NMMA groups compared with placebo on the basis of the Dunnett test or the Jonckheere–Terpstra test statistic (Z statistic = 0.186; two-sided P-value = 0.852). Percent change in MAP was also not statistically different in any group compared with placebo. Marked variability was observed in change in MAP at 2 h (Figure 3). Considerable down-titration of vasopressors occurred, most frequently in the two higher l-NMMA dose groups (Figure 4). There was no correlation between blood l-NMMA level and change in MAP.

Figure 3

Change in MAP 2 h after initiation of study drug bolus and infusion. Individual data points and median change in MAP are shown for each group. P values comparing the median change in each study group compared to placebo are as follows: P = 0.12, P = 0.71, P = 0.21 and P = 0.40. Mean change in MAP in mm Hg for each study group and P values comparing each treatment group against placebo are as follows: −0.4, 4.4 (P = 0.42), 1.8 (P = 0.69), −4.1 (P = 0.27) and 6.8 (P = 0.30). MAP = mean arterial pressure.

Figure 4

Proportion of patients in each study group with vasopressor down-titration prior to 2 h.

Secondary endpoints

MAP was also measured 15 min after study drug initiation, at which time vasopressor down-titration is unlikely to have occurred. Precise information on vasopressor dosing was not available at 15 min. Significant increases in MAP were noted for each l-NMMA group compared with placebo, except for the l-NMMA 1.0 mg/kg group, where a strong trend was observed (Figure 5). Changes in other haemodynamic parameters at 2 h, as well as change in blood glucose and total urine output at 24 h, were similar between groups (Table 4).

Figure 5

Change in MAP 15 min after initiation of study drug bolus and infusion. Individual data points and median change in MAP are shown for each group. P values comparing the median change in each study group compared to placebo are as follows: P = 0.013, P = 0.02, P = 0.012 and P < 0.001. Mean change in MAP in mm Hg for each study group and P values comparing each treatment group against placebo are as follows: −4.0, 5.8 (P = 0.02), 4.8 (P = 0.02), 5.1 (P = 0.07) and 11.6 (P < 0.001). MAP = mean arterial pressure.

View this table:
Table 4

Change in other haemodynamic parameters

Placebo (n = 20)l-NMMA 0.15 mg/kg (n = 15)l-NMMA 0.50 mg/kg (n = 15)l-NMMA 1.0 mg/kg (n = 15)l-NMMA 1.5 mg/kg (n = 14)
ΔPCWPa at 2 h0.3 ± 6.2−1.9 ± 5.0−1.9 ± 2.6−0.8 ± 4.0−1.0 ± 3.8
ΔSVRb at 2 h−58 ± 438126 ± 71550 ± 36595 ± 504206 ± 548
ΔPVRc at 2 h−5.6 ± 172.566.0 ± 244.2114.0 ± 135.196.6 ± 94.272.9 ± 151.6
ΔCId at 2 h0.1 ± 0.40.1 ± 0.40.0 ± 0.2−0.2 ± 0.40.0 ± 0.4
ΔCPe at 2 h0.1 ± 0.20.1 ± 0.20.0 ± 0.2−0.1 ± 0.20.0 ± 0.2
ΔGlucosef at 24 h−85.9 ± 99.8−46.6 ± 91.8−77.8 ± 103.8−65.0 ± 84.6−84.0 ± 169.9
Urine outputg over 24 h1774 ± 14542386 ± 16562375 ± 26062466 ± 17761579 ± 1111
  • All values are expressed as mean and standard deviation. All P-values are NS after adjustment for multiple comparisons.

  • aMean pulmonary capillary wedge pressure (PCWP) is used as an approximation of mean left atrial pressure.

  • bSVR = (mean arterial pressure−mean right atrial pressure×80)/CO.

  • cPVR = (mean pulmonary artery pressure−mean left atrial pressure×80)/CO.

  • dCI = cardiac output divided by body surface area.

  • eCardiac power (CP) = CI×MAP.

  • fSerum glucose measured in local lab.

  • gUrine measured at bedside (mL).


Analysis of adverse events (Table 5) revealed that l-NMMA was well tolerated and had a safety profile similar to placebo in this very high risk population. Importantly, exacerbation of RV failure, a problem identified with higher doses of l-NMMA in prior studies of septic shock,14 was not observed in this study. There were no cases of premature discontinuation of study drug infusion because of an adverse event.

View this table:
Table 5

Adverse events during study drug infusion; active study drug groups pooled

l-NMMA injection (n = 59)Placebo (n = 20)
Cardiovascular18 (30.5%)6 (30.0%)
Progressive cardiogenic shock9 (15.3%)3 (15.0%)
Mechanical1 (1.7%)0 (0%)
Arrhythmia4 (6.8%)1 (5.0%)
 Ventricular tachycardia/fibrillation6 (10.2%)2 (10.0%)
 Atrial fibrillation/flutter6 (10.2%)2 (10.0%)
 Pulseless electrical activity1 (1.7%)1 (5.0%)
Other cardiac3 (5.1%)1 (5.0%)
Gastrointestinal haemorrhage1 (1.7%)1 (5.0%)
Respiratory4 (6.8%)0 (0%)
Stroke0 (0%)0 (0%)
Renal and urinary disorder4 (6.8%)2 (10.0%)
Other6 (10.1%)5 (25.0%)


No significant differences in mortality were observed among any treatment groups at 30 days (Table 6). A pre-specified analysis by age revealed a rate of mortality much lower than projected among patients in the placebo group aged < 75 years. No association was observed between change in MAP by 2 h and survival in any treatment group or in the group as a whole.

View this table:
Table 6

All-cause mortality at 30 days (95% CI)

Placebol-NMMA 0.15 mg/kgl-NMMA 0.50 mg/kgl-NMMA 1.0 mg/kgl-NMMA 1.5 mg/kg
All patients (n = 20, 15, 15, 15, 14)35.0 (15.4, 59.2)53.3 (26.6, 78.7)53.3 (26.6, 78.7)26.7 (7.8, 55.1)28.6 (8.4, 58.1)
Age<75 years (n = 10, 12, 9, 10, 9)10.0 (0.3, 44.5)41.7 (15.2, 72.3)44.4 (13.7, 78.8)20.0 (2.2, 55.6)22.2 (2.8, 60.0)
Age≥75 years (n = 10, 3, 6, 5, 5)60.0 (26.2, 87.8)100 (29.2, 100)66.7 (22.3, 95.7)40.0 (5.3, 85.3)40.0 (5.3, 85.3)


We have demonstrated the feasibility of a randomized, multicentre international trial of a pharmacological agent in patients with AMI complicated by CS. However, in contrast to prior reports, we did not observe a significant change in the primary outcome measure of MAP at 2 h. There was no improvement in MAP relative to placebo at 2 h at any dose of l-NMMA studied, including the 1.0 mg/kg dose used in prior studies. Furthermore, we did not observe any association between blood l-NMMA level and change in MAP at 2 h. The observed lack of association between rise in MAP over the first 2 h and survival in our cohort suggests that early increases in MAP may be an inadequate surrogate for survival in CS complicating AMI. Finally, we have demonstrated the tolerability of, and provided safety data for, a 10-fold range of doses of l-NMMA bolus and 5 h infusion.

No untoward effects of l-NMMA on RV function or PVR at any dose were observed. A phase III trial of NOS inhibition in septic shock, utilizing doses of l-NMMA up to 13-fold higher than the maximum in our study, was stopped owing to excess mortality related to pulmonary hypertension and cardiac failure in the l-NMMA group.14

The results of SHOCK-2 contradict the findings of Cotter et al.,10 who reported significant improvement in BP, urine output, and other haemodynamic parameters, first in an 11-patient non-randomized study utilizing l-NMMA 1 mg/kg bolus and a 1 mg/kg/h 5 h infusion, and later in a small single-centre, randomized open-label pilot study (LINCS) of 30 patients, utilizing nitro-l-arginine-methyl-ester (L-NAME), a similar NOS inhibitor, also at 1 mg/kg bolus and 5 h infusion.11 The basis for these studies derives from experimental evidence that high levels of NO and NO-derived species cause a wide array of detrimental effects in both ischaemic and non-ischaemic myocardium,1519 pro-inflammatory effects, and reduction of catecholamine responsivity,20,21 as well as observations that NOS inhibition ameliorates myocardial ischaemia-reperfusion injury.22,23 In the randomized LINCS study, in which vasopressor therapies and fluids were managed by physicians who were not blinded to treatment, a significant survival benefit of NOS inhibition was suggested, with 30 day mortality of 37% in the l-NAME group and 70% in the control group.

Unfortunately, the current blinded, multicentre study did not replicate these encouraging haemodynamic and survival findings, which suggests l-NMMA may not be as efficacious as originally suggested. However, there may be other explanations for the divergent haemodynamic findings. The LINCS population had lower MAP than patients enrolled in SHOCK-2 (median on vasopressors and off IABP: 62 mmHg in LINCS and 71 mmHg in SHOCK-2). It is possible that l-NMMA has a larger effect on BP in patients with more severe haemodynamic derangements.

Also, in contrast to LINCS, in which doses of all vasoactive medications were maintained during l-NMMA infusion, considerable down-titration of vasopressor/inotropic agents occurred during the first 2 h of study drug administration to SHOCK-2 patients. In the present study, improvement in MAP in active drug groups was observed at 15 min, at which point down-titration of vasopressors was unlikely to have occurred. Facilitation of catecholamine responsivity by l-NMMA has been shown in a shock model,24 and this interaction could explain higher rates of vasopressor down-titration in higher dose l-NMMA groups in SHOCK-2. Decreases in vasopressor doses were presumably made in response to an improving circulatory state, and the distribution of changing vasopressor doses across the study drug dose range could indicate a dose-dependent effect on haemodynamics. The early rise in MAP with active study drug (seen at 15 min) suggests that excess NO does play a role in the hypotension associated with CS.

Study limitations

The small sample size limits conclusions about safety as well as efficacy of l-NMMA. The unexpectedly low mortality rate in the placebo group, rarely seen in persistent CS complicating AMI, suggests that the populations of earlier studies may not have been adequately replicated. Assessment of the effect of l-NMMA on haemodynamics was confounded by the differential use of vasopressors, possibly owing to haemodynamic effects of the active drug. Finally, any attempt to determine the most efficacious dose of l-NMMA is hampered by the fact that there is no known surrogate endpoint that correlates well with an effect of a pharmacological agent on mortality.


Although the potent haemodynamic effects of NOS inhibition in patients with CS suggested by prior studies have not been confirmed by SHOCK-2, this study has demonstrated the feasibility of a multicentre, randomized trial of an investigational agent in CS complicating AMI. Whether l-NMMA improves survival of patients with AMI and persistent CS can only be determined by an adequately powered study assessing all-cause mortality. The TRIUMPH study, utilizing an l-NMMA dose of 1.0 mg/kg bolus and 5 h infusion on the basis of evidence of prior studies and the tolerability demonstrated in this study, was designed to provide a definitive answer to this question.


This trial was funded by Arginox Pharmaceuticals (Redwood Shores, CA, USA) and supported in part by National Institutes of Health grant HL74702 to S.S.G.

Conflict of interest: Significant institutional research grants from Arginox: V.D., H.R.R., J.H.A., D.A.B., R.H., S.L.H., J.S.H.; modest research grant from Arginox: Z.V., J.E.P.; speaker honoraria from Arginox: K.R., M.F.; speaker honoraria from Datascope: V.D.; consultant honoraria from Datascope: J.S.H.; consultant/Arginox advisory board member: G.C., J.E.P.; significant research grant from Arginox to employer (DCRI): G.C.; employee of Arginox: D.H.; ownership interest in Arginox: D.H., S.S.G.; stock and other compensation from Arginox to employer (Cornell University): S.S.G. S.S.G. is also co-inventor of patents worldwide for the use of NOS inhibitors to treat hypotensive conditions such as CS.


View Abstract