Evaluation of lecithinized human recombinant super oxide dismutase as cardioprotectant in anthracycline-treated breast cancer patients

Abstract

Aim

Anthracycline-induced cardiotoxicity is (partly) mediated by free radical overload. A randomized study was performed in breast cancer patients to investigate whether free radical scavenger super oxide dismutase (SOD) protects against anthracycline-induced cardiotoxicity as measured by changes in echo, electrocardiography and an array of biomarkers.

Method and Results

Eighty female, chemotherapy-naïve breast cancer patients (median age 49, range 24–67 years) scheduled for four or five courses of adjuvant 3 weekly doxorubicin plus cyclophosphamide (AC) chemotherapy, were randomly assigned to receive 80 mg PC-SOD (human recombinant SOD bound to lecithin) or placebo, administered intravenously (i.v.) immediately prior to each AC course. The primary end point was protection against cardiac damage evaluated using echocardiography, QT assessments and a set of biochemical markers for myocardial function, oxidative stress and inflammation. Assessments were performed before and during each course of chemotherapy, and at 1, 4 and 9 months after completion of the chemotherapy regimen. In all patients cardiac effects such as increases in NT-proBNP concentration and prolongation of the QT_c interval were noticed. There were no differences between the PC-SOD and placebo-treated patients in systolic or diastolic cardiac function or for any other of the biomarkers used to assess the cardiac effects of anthracyclines.

Conclusion

PC-SOD at a dose of 80 mg i.v. is not cardioprotective in patients with breast carcinoma treated with anthracyclines.

What Is Already Known about This Subject

• Late occurrence of congestive heart failure is an important side effect of anthracycline therapy.

• The mechanism of anthracycline induced cardiotoxicity is not completely elucidated but the formation of free radical species seems to play an important role.

• Several free radical scavenging agents have been investigated, however with limited efficacy in clinical practice.

What This Study Adds

• In this study a free radical scavenger again failed to show a cardioprotective effect against anthracycline induced cardiotoxicity, which suggests that free radical overload is not the sole explanation for its cardiotoxicity.

• Several biomarkers mechanistically related to acute anthracycline induced myocardial damage were identified.

Introduction

Anthracyclines are widely used in treatment regimens of cancer, including breast cancer. Their use is hampered by the occurrence of irreversible cardiotoxicity which typically manifests as congestive heart failure (CHF) months to years after anthracycline exposure. It is primarily related to cumulative anthracycline dose and it seems that females are affected more often than males [1,2]. The incidence increases from 5% in patients receiving doses up to 400 mg m⁻² to 48% in patients receiving more than 700 mg m⁻² of doxorubicin [1]. Although less toxic analogues such as epi-doxorubicin have been developed, anthracycline-induced cardiotoxicity remains a clinical problem [3]. As the decline in ejection fraction and clinically manifest CHF usually become apparent relatively late after anthracycline therapy, it is difficult to assess the cardiotoxic effects of anthracyclines early. However, anthracycline cardiac toxicity has also been reported to occur after only single dose administration [1]. This suggests that it may be possible to use (bio)markers of cardiac effects due to anthracyclines occurring early and that may be predictive of the late toxicity. Indeed, several markers such as QT prolongation and changes in NT-proBNP and cardiac troponin concentrations have been suggested to be such early markers [4–7]. These markers can potentially also be used to assess the effects of putative protective strategies [4].

The mechanism of anthracycline-induced cardiotoxicity has not been fully elucidated, but formation of reactive oxidative species (ROS), such as the superoxide (O₂⁻) radical seem to play a major role [8]. Superoxide dismutase (SOD) is an important scavenger of these ROS and its use to prevent organ damage mediated by free radical overload has been investigated [8]. However, the currently existing therapies using exogenous SOD as a protectant have been limited by, for instance, its short half-life and low affinity for the cell membrane [9]. As a possible solution lecithinized SOD (PC-SOD) has been developed, which has a 100–200 fold higher affinity for the cell membrane and improved free radical scavenging properties [10]. Several animal models, including a rodent doxorubicin-induced cardiotoxicity model, showed that PC-SOD protected against free radical mediated injuries [11–20]. Early clinical studies in healthy subjects showed that a single intravenous (i.v.) dose of 80 mg PC-SOD resulted in increased SOD-activity in vivo for 16–24 h [21,22].

The efficacy of PC-SOD as a cardioprotective agent against anthracycline-induced cardiotoxicity was explored in an early phase II study utilizing serial echocardiography measurements, electrocardiography and a set of (bio)markers, reflecting myocardial function, oxidative stress and inflammation [4,7,23–28].

Methods

Patient population

This multicentre, randomized, placebo-controlled trial was performed in female patients with early stage breast cancer eligible for adjuvant doxorubicin and cyclophosphamide (AC) chemotherapy. Patients were scheduled to receive either four or five AC cycles according to national guidelines at that time. Prior or concomitant use of cardiotoxic medication was an exclusion criterion. Patients with distant metastases, a history of other malignant disease, a life expectancy of less than 1 year, pre-existing cardiovascular diseases, elevated transaminases above three times the upper limit of normal and patients in whom we were unable to obtain a good quality echocardiogram before study drug administration were excluded.

The institutional review board of Leiden University Medical Center (LUMC) approved the study protocol before inclusion of the first patient and complied with the principles of ICH-GCP, the Helsinki declaration and Dutch laws and regulations. All patients gave written informed consent before participation. The study is registered at http://www.controlled-trials.com, number ISRCTN56637853.

Study protocol

This study was coordinated by the Centre of Human Drug Research (CHDR) and carried out in five oncology centres in The Netherlands. After randomization (1:1 to 80 mg PC-SOD or placebo) of eligible patients baseline assessments were done and the patients started their scheduled chemotherapy (four or five courses) consisting of a combination of doxorubicin (60 mg m⁻² over approximately 30 min) and cyclophosphamide (600 mg m⁻² over approximately 30 min) administered i.v.. Patients were admitted to the hospital on the morning of each chemotherapy course. After baseline assessments were completed the patients received PC-SOD or placebo as a 1 h i.v. infusion and immediately thereafter, anti-emetics followed by AC. After discharge in the afternoon, a 24 h visit took place in the morning of the following day. A similar procedure was repeated during a maximum of four courses. Patients receiving five courses received the study drug at the third course but no measurements were done. The median volume loading for each course was 300 ml h⁻¹ (in total approximately 850 ml in 4 h). After completion of chemotherapy, follow-up visits took place at 1, 4 and 9 months.

Study medication

PC-SOD consists of an average of four molecules of a lecithin derivative covalently bound to the human derived CuZn-SOD, produced by genetic recombination using E.coli as a host cell [10]. The lecithinized product has 3 × 10³ U SOD-activity mg⁻¹. A single batch of the lyophilized formulation was used. The PC-SOD formulation consisted of 80 mg PC-SOD and 133 mg sucrose and the placebo formulation only consisted of sucrose. PC-SOD and placebo were prepared for use by dissolution in 5% mannitol diluted with distilled water. All study medication was prepared at the LUMC hospital pharmacy according to GMP regulations.

Outcome measures

Efficacy

The primary efficacy variables were systolic and diastolic function (left ventricular ejection fraction [LVEF], E : A ratio) measured with (rest) echocardiography. Further efficacy measurements included electrocardiography (ECG) (QT assessments) and blood sampling for biomarkers of cardiac function or damage (NT-proBNP, CK-MB and troponin T), inflammation (macrophage inhibiting protein 1 (MIP-1), high sensitivity C-reactive protein (hsCRP), tumour necrosis factor alfa (TNF-α) and soluble intercellular adhesion molecule-1 (sICAM) and oxidative stress (oxLDL, urinary biopyrrin and non-protein bound iron [NPBI]).

Echocardiography, ECGs and blood sampling for determination of NT-proBNP, CK-MB (mass) and troponin T concentrations were done at baseline and 1 (including NPBI), 4 and 9 months after a full chemotherapy regimen.

ECG and blood sampling (for all biomarkers) was done before and at 24 h after the start of each chemotherapy course. In addition ECG-recordings were made and blood was sampled for determination of CK-MB (mass), troponin T and TNF-α concentration at 4 h after the start of each chemotherapy course.

Safety

During the study period haematology and blood chemistry were frequently assessed. Glomerular filtration rate (GFR) was determined during each course from 24 h creatinine clearance and during the follow-up visits the MDRD formula was used [29]. In addition, at the first follow-up visit antibodies against PC-SOD were determined.

All (serious) adverse events ((S)AE) were monitored from inclusion until last follow-up and (S)AEs and concomitant medication were classified according to the World Health Organization Adverse Reaction Terminology and drug (WHOart and WHOdrug) classification system.

After completion of the last chemotherapy cycle for every 10th patient (until 60 patients were included), an interim safety report was reviewed by an independent Data Monitoring Committee (DMC). This report included all occurred (S)AE and laboratory safety data. After each report the DMC informed the principle investigator if, in their opinion, the data raised any safety concerns. The DMC was blinded during the whole study period, but could request emergency deblinding of (a part of) the data when deemed necessary.

Quality of life

Quality of life (QoL) was assessed using two validated questionnaires developed by the European Organization for Research and Treatment of Cancer (EORTC), QoL (EORTC QLQ-C30, version 3) and QoL (EORTC QLQ-BR23) [30]. QoL was assessed at baseline, during each course and 1, 4 and 9 months after completion of chemotherapy.

Pharmacokinetics

Blood was sampled for the determination of PC-SOD serum concentrations during the first and last course at baseline and directly, 4 h and 23 h after the end of the infusion of PC-SOD or placebo.

Echocardiography

Echocardiography was performed at two locations in the Netherlands, the department of cardiology of LUMC, Leiden and the department of cardiology of Maasstad Ziekenhuis, Rotterdam. The examinations were performed by a single echographist in each centre and all examinations were supervised by an experienced cardiologist. To exclude inter-observer variability all echo assessments for each individual patient were done at one centre.

The investigations consisted of routine imaging, M-mode imaging for measurement of left ventricular end-diastolic and end-systolic wall thickness (septum, posterior wall), fractional shortening and left ventricular ejection fraction (LVEF, calculated according to Teichholz) [31]. Measurements were made from the parasternal long axis (or short axis) view. The ratio of early rapid ventricular filling over atrial assisted filling (E : A ratio) was measured using pulsed-wave Doppler. Regional systolic function was evaluated with visual assessment of wall motion (and wall motion score index, WMSI) according to the 16-segment model. The examinations were performed using a GE vivid-7 echocardiograph equipped with pulsed-wave Doppler at the LUMC and using a Hewlett-Packard HP 5500 with a S3 probe at the Maasstad Ziekenhuis.

ECG recordings and analysis

For each patient 5 min ECG recording were made using the CardioPerfect device (Welch Allyn, Delft, The Netherlands). ECG recordings were analyzed after fiducial segment averaging (FSA) to obtain heart rate, and QT interval. This analysis was done using Intraval (Advanced Medical Systems, Maasdam, the Netherlands) [20].

For the analyses correction of the QT interval for heart rate was done using Bazett's formula (QT_cB = QT/√ (RR)), Fredericia's cubic root QT_cF = QT*(1/RR)^1/3 and using the linear correction method according to Framingham Heart Study (QT_cL = QT + 0.154*(1 – RR)).

Assays

Samples were assayed for NT-proBNP, troponin T and CK-MB (mass) and NPBI at the Central Clinical Chemical laboratory (CKCL) of LUMC. Lower limits of detection (inter- and intra-assay variability) were 5 ng l⁻¹ (<5.8%), 0.01 ng ml⁻¹ (<2.5%), 0.01 μg ml⁻¹ (<5.6%) and 0.01 μmol l⁻¹ (<9.2) for NT-proBNP, cTnT, CK-MB (mass) and NPBI, respectively. Assays for TNF-α, hsCRP, sICAM, MIP-1α, oxLDL and urinary bioppyrin were performed at the Netherlands Organization for Applied Scientific Research (TNO). The lower limits of detection of the assays (inter- and intra-assay variability) were 0.12 pg ml⁻¹ (<12.5%), 0.1 μg l⁻¹ (<10%), 0.35 ng ml⁻¹ (<12.5%), 10 pg ml⁻¹ (<10%), 1 mU l⁻¹ (<7.5%) and 0.1 U l⁻¹ (<12.5%) for TNFα, hsCRP, sICAM, MIP-1α, oxLDL and urinary biopyrrin, respectively.

Serum PC-SOD concentrations were measured using an enzyme linked immunosorbent assay (ELISA), consisting of an antibody against human Cu, Zn-SOD, and a second antibody against human Cu, Zn-SOD conjugated with horseradish peroxidase. The assay has a lower limit of quantification of 626 ng ml⁻¹ and the coefficients of variation did not exceed 7.9% which was observed for the lower concentrations. Antibody formation against PC-SOD was measured by quantification of specific IgE, IgG and IgM titres as described previously [21].

Statistical analyses

Power

As both the incidence of sub-clinical cardiotoxicity and the treatment effects were unknown, the power calculation of the study has been performed using a number of assumptions: (i) the incidence of subclinical cardiomyopathy in patients is 33% and (ii) animal experiments suggest that PC-SOD treatment prevents cardiomyopathy in 100% of the cases. It has therefore been estimated that the protection by PC-SOD will reduce the incidence of cardiomyopathy from 33% to 5.5% in the patients. In order to be able to demonstrate this treatment effect (power = 80%. two-sided test. P < 0.05) a total of 72 patients is required. Second an exploratory power calculation on the biomarkers has been performed. These showed that this study has 80% power to detect (two- sided test. P < 0.05) a difference in NT-proBNP concentrations between the groups of approximately 53%.

Efficacy and safety population

Eighty female breast cancer patients were randomized to PC-SOD or placebo. After randomization one patient was excluded because of an abnormal echocardiography at baseline. During the trial eight patients were replaced. Four patients dropped out (two patient's request, two because of discontinuation of chemotherapy due to extreme nausea) and four patients had incomplete echocardiographic assessments (two due to logistic problems, two due to equipment failure). Data of replaced patients are used in both safety and efficacy analyses. Two patients were excluded from the efficacy analyses because of anomalies in PK results. Hence the safety population and efficacy population consisted of 79 and 77 patients, respectively (Table 1).

Table 1.

Subjects disposition

Group	n
Number of subjects screened	82
Number of subjects suitable after screening	80
Number of subjects randomized	80
Number of subjects treated	79
Number of subjects completed	76
Number of subjects in safety population	79
Number of subjects in efficacy population	77

Treatment effects

First, for the ECG parameters and the biomarkers of inflammation, oxidative stress and cardiac injury, for each course the difference between the 24 h and the baseline measurement was calculated and this series of four differences was compared between placebo and PC-SOD treatment. This short term effect was analyzed using a mixed model analysis of variance (SAS proc mixed) with visit as repeated factor within patient, treatment (PC-SOD/placebo), group (four or five courses) and treatment by group, treatment by time, group by time and treatment by group by time as fixed effects. For ECG parameters, TNF-α and CK-MB (mass) the difference between the measurement at 4 h and baseline at the occasion was analyzed the same way.

Second, the baseline measurements of each course, except the first, and the follow-up measurements were compared between placebo and PC-SOD. The long term treatment effect for the echocardiographic, ECG parameters and CK-MB and NT-proBNP was analyzed using a mixed model analysis of variance (SAS proc mixed) with visit (occasion) as repeated factor within patient, treatment, group, treatment by time, treatment by group, time by group and treatment by time by group as fixed effects. The baseline value of the first course was included as covariate.

Time effects

To assess the 4 and 24 h difference from baseline within a course, the estimated differences from baseline were compared with 0 (no difference from baseline) within the first treatment mixed model for the ECG and biomarker parameters. The estimated difference between the course 1 baseline and follow-up measurements (long term time effects) for the echocardiographic, ECG parameters, CK-MB and NT-proBNP were compared with 0 (no difference from baseline) within the second treatment mixed model.

Pharmacokinetics

Compartmental pharmacokinetic analysis was performed using nonmem Version VI software (GloboMax LLC, Hanover, MD, USA). Maximum serum concentration and half-life are reported.

Additional (sub-group analyses)

All data were also analyzed excluding patients who received trastuzumab or left-sided radiotherapy as concomitant therapy.

All statistical analyses were performed using SAS for windows V9.1.2 (SAS Institute, Inc., Cary, NC, USA).

Results

Baseline characteristics

The median age of the 79 patients who received at least one dose of PC-SOD and AC chemotherapy was 49 years (range 24–67 years). The median (min−max) number of courses was 4 (1 to 5) and 4 (2 to 5) for placebo and PC-SOD, combined with AC chemotherapy, respectively (Table 2).

Table 2.

Baseline demographics and clinical characteristics

	Placebo (n = 40)	PC-SOD (n = 39)
Age, BMI and cumulative doxorubicin dose; median (range)
Age (years)	47.0 (30–66)	50.0 (24–67)
BMI (kg m−2)	24.0 (19.4–37.7)	25.5 (19.6–38.5)
Cumulative doxorubicin dose (mg m−2)	240 (60–300)	240 (120–300)
Echocardiography; mean (SD)
LVEF (%)	66 ± 6	64 ± 7
E : A ratio	1.06 ± 0.28	1.08 ± 0.28
Haemoglobin, renal function and cardiac (bio)markers; mean (SD)
Haemoglobin (mmol l−1)	7.7 ± 0.7	8.0 ± 0.7
GFR (ml min−1)	100.5 ± 23.9	106.3 ± 37.4
NT-proBNP (ng l−1)	71 ± 46	84 ± 9 2
CK-MB mass (mg l−1)	1.7 ± 0.9	1.7 ± 0.5
QTcL (ms)	426 ± 12.7	433 ± 24.0
Number of courses*; n (%)
1	2 (3%)	0 (0%)
2	0 (0%)	1 (1%)
3	2 (3%)	0 (0%)
4	17 (22%)	20 (25%)
5	19 (24%)	18 (23%)
Adjuvant therapy; n (%)
Hormonal therapy	24 (60%)	26 (67%)
Trastuzumab	0 (0%)	5 (13%)
Docetaxel	0 (0%)	4 (10%)
Gosereline	2 (5%)	3 (8%)
Radiotherapy; n (%)
Right
Prior to chemotherapy	3 (8%)	3 (8%)
After cessation of chemotherapy	3 (10%)	9 (23%)
Left
Prior to chemotherapy	4 (10%)	4 (10%)
After cessation of chemotherapy	6 (15%)	9 (22%)

^*Number of courses doxorubicin, cyclophosphamide (all patients received PC-SOD or placebo prior to their chemotherapy courses). No statistical comparison was done at baseline, as baseline values were included as covariates.

Safety

There were no clinically relevant findings related to PC-SOD treatment on clinical laboratory measurements, vital signs or ECG findings. GFR was stable during the study period. The AE pattern did not differ among treatment groups (Table 3) and the majority of AEs could be attributed to the chemotherapeutics and were mild to moderate in intensity. One adverse event, severe dyspepsia requiring hospitalization, was considered possibly related to PC-SOD administration. Antibodies against PC-SOD were not detected in any of the patients

Table 3.

Summary of adverse events

Adverse event	Placebo (n = 40)		PC-SOD (n = 39)
	Number of patients	%	Number of patients	%
General
Fatigue	30	75	30	77
Malaise	6	15	8	21
Hot flushes	15	38	16	41
Surgery related AEs	7	18	8	21
Change of taste	11	28	8	21
Central nervous system disorders
Headache	26	65	18	46
Dizziness	8	20	8	21
Gastro-intestinal system disorders
Nausea	32	80	29	74
Constipation	11	28	11	28
Dyspepsia	7	18	13	33
Mucositis	5	13	9	23
Psychiatric disorders*	10	25	11	28
Respiratory system disorders
Respiratory tract infections	13	33	17	44
Vision disorders
(Kerato-) conjunctivitis	18	43	18	46

Note: Adverse events occurring during chemotherapy in more than 20 patients in one of the two treatment groups.

^*Nervousness, emotional lability, anxiety, agitation, insomnia, impaired concentration, abnormal thinking, depression and hallucinations.

Efficacy

Long term effects

Echocardiography

LVEF (±SD) and E : A ratio (±SD) were 66.6% ± 6.0, 1.06 ± 0.28 and 64.2% ± 6.8, 1.08 ± 0.28 at baseline, in patients receiving placebo and PC-SOD, respectively and the overall decline (95% confidence interval (CI) between brackets) was −1% (−2, 1%), −0.0 (−0.0, 0.0%) and −2% (−3, −0%), −0.04 (−0.0, 0.0%) during the study (Figure 1, Table 4).

Figure 1.

Mean left ventricular ejection fraction (A) and E : A ratio (B) for PC-SOD (open circles) and placebo (closed circles) and 95% CIs (PC-SOD down, placebo up) at baseline and 1, 4 and 9 months post-chemotherapy

Table 4.

Long term effects

	Overall change over course baseline and follow-up		Difference between treatments	95% CI	P-value
	Placebo	PC-SOD
Echocardiographic parameters
Left ventricular ejection fraction (%)	−1	−2	−1	−3, 1	0.31
E : A ratio	−0.0	−0.0	−0.0	−0.1, 0.0	0.48
Biomarkers of myocardial injury
NT-proBNP (ng l−1, % change)	32.0*	14.2	−13.5	−30.9, 8.2	0.20
CK-MB mass (mg l−1, % change)	−5.7	−7.6*	−2.0	−11.7, 8.8	0.70
Electrocardiography
Heart rate (beats min−1)	0.6	4.0*	3.4	0.4, 6.5	0.03
QT interval (ms)	6.8*	3.6	−3.1	−11.6, 5.4	0.46
QTcB interval (ms, corrected Bazett)	8.8*	16.2*	7.4	1.9, 12.9	<0.01
QTcL interval (ms, corrected Framingham)	7.5*	10.8*	3.3	−2.1, 8.7	0.23

^*Significant change from baseline (P < 0.05).

Differences (95% CI between brackets) between PC-SOD and placebo on LVEF and E : A ratio were −1% (−3, 1%) and −0.0 (−0.1, 0.0%), respectively.

WMSI did not change significantly during the trial and no differences between treatments were observed.

Biomarkers – myocardial injury

During courses and follow-up the overall change (percentage change, 95% CI between brackets) of NT-proBNP and CK-MB was 32.0% (12.8, 54.5), −5.7% (−12.4, 1.4%) and 14.2% (−2.6, 33.8%), −7.6% (−14.2,−0.5%) in patients receiving placebo and PC-SOD, respectively. (Figure 2, Table 4)

Figure 2.

Mean NT-proBNP (A) concentrations and QT_c (B) linearly corrected for heart rate according to Framingham, during chemotherapy and follow-up for PC-SOD (open circles) and placebo (closed circles) and 95% CIs (PC-SOD up, placebo down) during the course and 1, 4, 9 months post-chemotherapy

The differences between PC-SOD and placebo were (percentage change, 95% CI between brackets) −13.5% (−30.9, 8.2%) and −2.0% (−11.7, 8.8%) for NT-proBNP and CK-MB, respectively.

During the follow-up period in 10 patients (six PC-SOD, four placebo) detectable (although not pathological elevated) troponin levels were present.

Electrocardiography

Heart rate and QT_c interval (corrected using a linear method) increased during courses and follow-up by (mean change, 95% CI between brackets) 0.6 beats min⁻¹ (−1.5 to 2.8 beats min⁻¹), 5 ms (3.8 to 11.3 ms) and 4.0 beats min⁻¹ (1.9 to 6.2 beats min⁻¹), 10.8 ms (7.0 to 14.7 ms) in patients with placebo or PC-SOD, respectively. (Figure 2, Table 4).

The differences between PC-SOD and placebo for heart rate, QT interval, corrected QT interval (Bazett) and corrected QT interval (linear) were (mean change, 95% CI between brackets) 3.4 beats min⁻¹ (0.4 to 6.5 beats min⁻¹), −3.1 ms (−11.6 to 5.4 ms), 7.4 ms (1.9 to 12.9 ms) and 3.3 ms (−2.1 to 8.7 ms), respectively.

Short term effects

Oxidative stress

Urinary biopyrrin increased (percentage change, 95% CI between brackets) in the placebo group only, although this effect was not present for each individual course. Change at 24 h was 13.0% (0.8, 26.7%) and 3.4% (−7.9, 16.0%) in patients receiving placeboand PC-SOD, respectively. While oxLDL and NPBI concentrations did not change significantly between baseline and at 24 h, the difference in percentage change (95% CI between brackets) between PC-SOD and placebo was 10.3% (−20.5, 52.9%), 6.2% (0.2, 12.5%) and −8.5% (−22.2, 7.6%) for urinary biopyrrin, oxLDL and NPBI, respectively (Table 5).

Table 5.

Short term effects

	Overall change from course baseline (at 24 h)		Difference between treatments	95% CI	P value
	Placebo	PC-SOD
Biomarkers of oxidative stress
Urinary biopyrrin (μmol g−1 creatinine, % change)	13.0*	3.4	10.3	−20.5, 52.9	0.55
OxLDL (mU l−1, % change)	−3.0	3.0	6.2	0.2, 12.5	0.04
NPBI (μmol l−1, % change)	15.9	5.1	−8.5	−22.2, 7.6	0.28
Biomarkers of myocardial injury
NT-proBNP (ng l−1, % change)	199.8*	263.8*	21.4	−3.9, 53.3	0.10
CK-MB mass (mg l−1, % change)	8.2*	10.3*	2.0	−8.1, 13.1	0.71
Biomarkers of inflammation
hsCRP (μg l−1, % change)	−2.0	−4.8	−2.9	−19.0, 16.5	0.75
sICAM-1 (ng l−1, % change)	−1.1	−1.8*	−0.7	−2.7, 1.4	0.50
TNF-α (pg l−1, % change)	−23.2*	−27.2*	−5.3	−18.1, 9.6	0.46
MIP-1α (pg ml−1, % change)	−46.9*	−48.6*	−3.3	−45.7, 72.1	0.91
Electrocardiography
Heart rate (beats min−1)	−1.2	3.7*	4.9	2.2, 7.6	<0.001
QT interval (ms)	16.3*	5.3*	−11.0	−18.2, −3.9	0.003
QTcB interval (ms)	14.2*	17.1*	3.0	−2.5, 8.4	0.23
QTcL interval (ms)	14.2*	11.5*	−2.7	−7.4, 1.9	0.24

^*Significant change from baseline (P < 0.05).

Myocardial injury

The overall increment at 24 h post-dose for NT-proBNP and CK-MB (mass) (mean change, 95% CI between brackets) was 199.8% (154.6, 253.0%), 8.2% (0.6, 16.4%) and 263.8% (207.9, 329.7%), 10.3% (2.5, 18.8%) in patients receiving placebo and PC-SOD, respectively.

The difference (mean change, 95% CI between brackets) at 24 h post-dose between PC-SOD and placebo was 21.4% (−3.9, 53.3%) and 2.0% (−8.1, 13.1%) for NT-proBNP and CK-MB, respectively (Table 5).

Inflammation

hsCRP and sICAM did not change markedly, while TNF-α and MIP-1α declined with (mean change, 95% CI between brackets) −23.2% (−30.5, −15.1%), −46.9% (−64.5, −20.4%) and −27.2% (−34.5, −19.1%), −48.6% (−65.9, −22.5%) in patients receiving placebo or PC-SOD, respectively. For TNF-α this effect was already present at 4 h post-dose. At 24 h the difference (mean change, 95% CI between brackets) between PC-SOD and placebo for hsCRP, sICAM-1, TNF-α and MIP-1α was −2.9% (−19.0, 16.5%), −0.7% (−2.7, 1.4%), −5.3% (−18.1, 9.6%) and −3.3% (−45.7, 72.1%), respectively (Table 5).

Electrocardiography

Heart rate showed a small increment at 24 h after each course in the PC-SOD arm, changes were (mean change, 95% CI between brackets) −1.2 beats min⁻¹ (−3.2, 0.7 beats min⁻¹) and 3.6 beats min⁻¹ (1.8, 5.5 beats min⁻¹) in patients receiving placebo and PC-SOD, respectively. After each course the (corrected) QT interval was prolonged at 4 h post-dose and increased further at 24 h post-dose. Overall prolongation of the QT_c interval (using a linear correction method) at 24 h post-dose was (mean change, 95% CI between brackets) 12.4 ms (8.8, 15.9 ms) and 9.8 ms (6.4, 13.2 ms) in patients receiving placebo and PC-SOD, respectively.

The difference between PC-SOD and placebo at 24 h after each chemotherapy cycle in heart rate, QT interval, corrected QT interval (Bazett), QT interval (linear) was (mean change, 95% CI between brackets) 4.9 beats min⁻¹ (2.2, 7.6 beats min⁻¹), −11.0 ms (−18.2, −3.9 ms), 3.0 ms (−2.5, 8.4 ms) and −2.7 ms (−7.4, 1.9 ms), respectively (Table 5).

Number of courses and other adjuvant therapy

It was also analyzed whether other (potentially cardiotoxic) adjuvant therapy or the number courses influenced our results. As all analyses showed comparable results; only the full dataset was reported.

Quality of life

In both treatment groups similar effects (decline) on QoL during the chemotherapy were observed (data not presented).

Pharmacokinetics

Maximum serum concentrations were reached within 1 h and amounted to 32.4 mg l⁻¹ (SD 11.9) and 31.4 mg l⁻¹ (SD 12.1) for the first and last visit, respectively. The estimated half-life was approximately 20 h.

Discussion

The main outcome of this study was that PC-SOD did not show a protective effect on cardiotoxicity, as no differences in echographic systolic and diastolic function were observed. Although NT-proBNP concentrations and QT_c interval changed significantly in in all breast cancer patients undergoing AC chemotherapy, no effect of PC-SOD was seen on these parameters, Also, any of the array of the other biomarkers assessed did not show a clinically significant change during or following chemotherapy and were also not affected by PC-SOD.

Safety analyses did not show any unfavourable effects of PC-SOD at the administered dose, as laboratory assessments and AE patterns were similar between treatments. Only one SAE (hospital admission for severe pyrosis) was considered possibly related to the administration of the study medication, although administration of PC-SOD is not associated with an increased incidence of gastro-intestinal complaints.

The lack of a cardioprotective effect of PC-SOD at a dose of 80 mg i.v. on any of the markers of anthracycline-induced cardiotoxicity in chemo-naïve breast cancer patients may be explained by a lack of efficacy of PC-SOD at the dose used.

The negative findings in this study are in keeping with the results of several other studies, showing that exogenously administered free radical scavengers are not able to protect against anthracycline-induced cardiotoxicity and add to the increasing knowledge that free radical mediated injury is only partly involved in the pathogenesis of the cardiotoxicity of anthracyclines [32–34]. In addition, we were not able to demonstrate the occurrence of oxidative stress in vivo, as none of the biomarkers for oxidative stress changed after doxorubicin infusion.

Another reason for the lack of effects of PC-SOD (and maybe of free radical scavenging agents in general) could be that the therapeutic window of these agents seems to be narrow. This involves the observation that in animals a bell-shaped dose−response curve (higher doses of SOD showed less protection) is present after administration of (PC-) SOD [11,35–39]. If such a bell-shaped curve is also present in humans, this could implicate that in this study the correct dosage was not used. Although several mechanisms could be responsible for this bell-shaped effect curve, the most plausible explanation is (PC-) SOD causing excess ROS formation. In particular this concerns formation of H₂O₂ which has been shown tobe capable of inducing apoptosis in cardiomyocytes [8,37,40–43].

Independent of the explanation of the failure of free radical scavenging agents as protective agents against anthracycline-induced cardiotoxicity, our study re-emphasizes the necessity to identify other strategies to reduce the risk of anthracycline-induced CHF.

We considered the possibility that the administered doses of anthracyclines did not induce sufficient myocardial damage to detect any prophylactic effect of PC-SOD, as LVEF and E : A ratio did not change markedly. However, the profound changes in NT-proBNP concentration and (corrected) QT interval, indicate that all patients indeed experienced some (subclinical) cardiotoxicity, as both markers are associated with the occurrence of anthracycline induced cardiac failure and an adverse outcome [44].

A limitation of our study is that although the echo and electrocardiographic and biochemical endpoints used in this study are well established markers of (anthracycline induced) cardiac damage and functional impairment, oxidative stress and inflammation, the study was not designed to detect differences in cardiac mortality or the occurrence of clinical CHF. Furthermore, some patients received additional potentially cardiotoxic treatments such as trastuzumab and/or radiotherapy. However, re-analysis of the data excluding these patients did not result in different findings, so, this did not influence our conclusions.

In conclusion, we showed that i.v. administration of 80 mg PC-SOD prior to each chemotherapy course was not efficacious as a protective agent against anthracycline-induced cardiotoxicity, as evaluated by echocardiography, electrocardiography and a comprehensive array of biomarkers of myocardial damage, inflammation and oxidative stress, in female breast cancer patients treated with a combination of cyclophosphamide and doxorubicin for early stage breast cancer.

Role of the funding source

The funding source was involved in the design of the study. Data collection and analyses were performed by CHDR. All authors had access to all study data and had final responsibility for the decision to submit for publication.

Competing Interests

All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare FB, AC, MK and JB are employees of the CHDR, that was financially supported by LTT Bio-Pharma for the submitted work. JS and YM were employees of LTT Bio-Pharma during the study.