OUP user menu

Circulating microparticles as indicators of peripartum cardiomyopathy

Katrin Walenta, Viktoria Schwarz, Stephan H. Schirmer, Ingrid Kindermann, Erik B Friedrich, Erich Franz Solomayer, Karen Sliwa, Saida Labidi, Denise Hilfiker-Kleiner, Michael Böhm
DOI: http://dx.doi.org/10.1093/eurheartj/ehr485 1469-1479 First published online: 3 February 2012


Aims Peripartum cardiomyopathy (PPCM) is associated with high mortality and morbidity. Endothelial damage involving cathepsin-D to form a 16 kDa prolactin (PRL) peptide is pathogenetically relevant. Inhibiting PRL peptide with bromocriptine has yielded promising results. We investigated whether microparticles (MPs) can be quantified in serum as markers for diagnosis and treatment effects in PPCM.

Methods and results Patients with PPCM were compared with age-matched healthy post-partum women (PPCTR), healthy pregnant women (PCTR), healthy non-pregnant women (NPCTR), patients with ischaemic cardiomyopathy (ICM), patients with stable coronary artery disease (CAD) and healthy controls (HCTR). Peripartum cardiomyopathy treated with bromocriptine (PPCM-BR) and with PPCM without bromocriptine-treatment as control (PPCM-BRCTR) were compared. Microparticles were determined by flow cytometry. Endothelial MPs (EMPs) were elevated in PPCM compared with PPCTR, PCTR, and NPCTR, each P< 0.001. They were significantly elevated compared with ICM, CAD, and HCTR (P< 0.001). Pregnancy (PCTR) exhibited only slight increases vs. ICM, CAD, NPCTR, and HCTR. The increase in PPCM was due to an increase of activated but not apoptotic EMPs. Platelet-derived microparticles were highly increased in PPCM compared with ICM (P< 0.001) but 9.3 ± 4.4-fold compared with CAD (P< 0.001). In NPCTR (P< 0.001) compared with NPCTR, the increase was 5.9 ± 1.7-fold (P< 0.001). Microparticles generated from monocytes (MMPs) were increased 2.4 ± 1.8-fold in PPCM compared with PCTR (P< 0.001) and 4.8 ± 3.6-fold compared with CAD (P< 0.001), whereas leucocyte MPs (LMPs) were not significantly elevated. Endothelial microparticles were significantly reduced in PPCM treated additionally with bromocriptine compared with PPCM treated only with heart failure therapy (P< 0.001).

Conclusion Microparticle profiles may in long-term distinguish PPCM from normal pregnancy, heart failure, and vascular diseases and might be a diagnostic marker related to the pathomechanism of PPCM.

  • Peripartum cardiomyopathy
  • Heart failure
  • Vascular damage
  • Microparticles


Peripartum cardiomyopathy (PPCM) is associated with severe heart failure and has an incidence of 1 : 2289–4000 life-births in the USA,1 1 : 1000 in South Africa2, and 1 : 299 in Haiti.3 Maternal mortality and normalization of left ventricular (LV) ejection fraction occurs only rarely with high incidence of developing heart failure despite therapy.13 Transgenic mice defective in STAT3 involved in protection from oxidative stress develop PPCM.4,5 Consecutive lack of anti-oxidative enzymes, protective in the post-natal heart, induces increased oxidative stress which in turn enhances cathepsin-D (CD) activity leading to proteolytic cleavage of prolactin (PRL) into its detrimental 16 kDa form producing endothelial cell apoptosis, capillary dissociation, and vasoconstriction.4 Inhibition of PRL with bromocriptine prevented the formation of a PPCM in animal models6,7 and led to improvement of LV function in patients.6 However, diagnosis of PPCM is often missed at early stages of the life-threatening disease and a marker as diagnostic test is lacking to early initiate this potentially life-saving therapy.

Disruption of endothelial integrity is well known to represent a crucial event in the initiation and development of cardiovascular diseases.8,9 It is now well established that microparticles (MPs), vesicles released from cellular membranes during cell activation/apoptosis,10 may play a pivotal role as bioactive molecules.1116 Endothelial MPs (EMPs) could directly demonstrate endothelial damage in PPCM. Furthermore, platelet-derived (PMPs),17 monocyte-derived (MMPs) and leucocyte-derived MPs (LMPs)18 could indicate the subsequent detrimental events like clot formation and inflammatory activation.15,17,18 In previous studies, MPs have been suggested to be a sign of severity in preeclampsia19 and MPs gained from women with preeclampsia furthermore elicit an intense vascular wall inflammation.20 We speculated that a typical MP profile could be of diagnostic importance in PPCM. Therefore, we investigated circulating MPs in PPCM in comparison to patients with known cardiovascular disease, such as acute heart failure with ischaemic cardiomyopathy (ICM) and stable coronary heart disease (CAD), on the one hand, and to healthy post-partum (PPCTR) and healthy pregnant (PCTR) women, on the other hand, in order to distinguish between physiological changes during pregnancy and post-partum. Besides this MPs were compared with non-pregnant (NPCTR) women as healthy age-matched controls.


Blood samples were obtained from PPCM patients (NYHA functional class III or IV) treated in line with the German registry for PPCM. As controls, age-matched post-partum women (PPCTR) were enclosed in Hannover, Germany. Age- and pregnancy-matched healthy women (PCTR), healthy non-pregnant women (NPCTR), participants with symptomatic ICM with clinical signs of heart failure (ICM), stable coronary artery disease (CAD) and age-matched to ICM and CAD healthy controls (HCTR) were included in Homburg, Germany. In total, all 20 patients in the bromocriptine-study were included at the Chris Hani Baragwanath Hospital. Patients were referred from local clinics, secondary hospitals, and the Department of Obstetrics at the Chris Hani Baragwanath Hospital. All patients were included and randomized with a computer-generated randomization list (blinded) within 24 h of diagnosis. All patients received treatment with the same heart failure therapy (furosemide and enalapril). The 10 patients randomized to standard therapy (PPCM-BRCTR group) were treated as outlined above. The 10 patients randomized to standard therapy plus bromocriptine (PPCM-BR) received bromocriptine 2.5 twice daily for 2 weeks followed by 2.5 mg daily for 6 weeks in addition to standard heart failure therapy. After informed consent, 10 mL peripheral venous blood was sampled from each of the 158 enrolled subjects. Measurements considered one point of time and no comparison to a follow-up. Microparticles were isolated from women with PPCM (n= 24, aged 33 ± 7 years, fractional shortening (FS) 13 ± 5%), age-matched PPCTR (n= 18, aged 30 ± 7 years), PCTR (n= 18, aged 31 ± 6 years, FS: 37 ± 2%), age-matched healthy NPCTR (n= 19, aged 30 ± 7 years, FS: 35 ± 4%), patients with ICM (n= 17, aged 71 ± 8 years, FS: 23 ± 3%) and CAD (n= 20, aged 71 ± 8 years, FS: 31 ± 3%) and age-matched HCTR (n= 22, aged 71 ± 7 years, FS: 38 ± 7%). In the subsequent investigation, patients with PPCM from South Africa were enclosed for the analysis of MPs in patients with or without bromocriptine therapy administered immediately after delivery (with therapy: PPCM-BR, n= 10, aged 24 ± 6 years, ejection fraction (EF) 27 ± 8%, without bromocriptine therapy: PPCM-BRCTR, n= 10, aged 28 ± 10 years, EF: 26 ± 8%). Demographic and clinical data are summarized in Table 1. The study was approved by the appropriate ethics committee. All patients gave written informed consent to include their data in the study.

View this table:
Table 1

Basic characteristics of patients with peripartum cardiomyopathy, age-matched healthy pregnant, healthy post-partum, and healthy non-pregnant women, patients with ischaemic cardiomyopathy, stable coronary artery disease, and age-matched healthy controls

PPCM (n= 24)PPCTR (n= 18)PCTR (n= 18)NPCTR (n= 19)ICM (n= 17)CAD (n= 20)HCTR (n= 22)PPCM-BR (n= 10)PPCM-BRCTR (n= 10)
Clinical parameters
 Mean age (years)±SD33 ± 730 ± 731 ± 630 ± 771 ± 871 ± 871 ± 724 ± 628 ± 10
 Gender (m:w)0 : 140 : 100 : 140 : 148 : 49 : 59 : 40 : 100 : 10
Echocardiographic parameters
 LVEDD (mm)n.d.n.d.41 ± 642 ± 460 ± 350 ± 653 ± 555 ± 1059 ± 5
 LVESD (mm)n.d.n.d.26 ± 427 ± 346 ± 232 ± 937 ± 446 ± 952 ± 6
 IVSD (mm)n.d.n.d.11 ± 211 ± 113 ± 216 ± 315 ± 2n.d.n.d.
 IVDD (mm)n.d.n.d.9 ± 19 ± 111 ± 113 ± 211 ± 1n.d.n.d.
 LVPWD (mm)n.d.n.d.10 ± 110 ± 110 ± 112 ± 111 ± 2n.d.n.d.
 LVPS (mm)n.d.n.d.11 ± 112 ± 115 ± 216 ± 512 ± 5n.d.n.d.
 FS (%)13 ± 5n.d.37 ± 235 ± 423 ± 331 ± 338 ± 7n.d.n.d.
 LVEF (%)26 ± 12n.d.n.d.n.d.40 ± 674 ± 1283 ± 927 ± 826 ± 8
Mitral regurgitation (n)
 Grade I6n.d.1164312
 Grade II13n.d.0042277
 Grade III5n.d.0020021
 Grade IV0n.d.0000000
Laboratory parameters:
 CK (U/L)283 ± 4200 ± 00 ± 0110 ± 1446 ± 1973 ± 679 ± 5n.d.n.d.
 CK-M (U/L)30 ± 100 ± 00 ± 0<1414 ± 215 ± 2<14n.d.n.d.
 ASAT (U/L)120 ± 2840 ± 00 ± 011 ± 556 ± 6140 ± 2122 ± 931 ± 1250 ± 27
 ALAT (U/L)104 ± 1780 ± 00 ± 020 ± 528 ± 1043 ± 2317 ± 544 ± 5657 ± 49
 LDH (U/L)501 ± 4530 ± 00 ± 0120 ± 45251 ± 31193 ± 35193 ± 62n.d.n.d.
 Creatinine (mg/dL)1.12 ± 0.810 ± 00 ± 01 ± 03 ± 51 ± 01 ± 00.7 ± 1.00.6 ± 1.1
 Troponin (ng/mL)0.6 ± 1.0<0.01<0.01<0.011.2 ± 0.06<0.01<0.010.8 ± 1.00.7 ± 0.6
 NT-pro-BNP (pg/mL)5860 ± 62740 ± 00 ± 00 ± 012 769 ± 4200 ± 01077 ± 76910 310 ± 16 9868992 ± 10 644
 CRP (mg/mL)46 ± 720 ± 00 ± 04 ± 12 ± 193 ± 24 ± 222 ± 3718 ± 20
 Fribrinogen (mg/dL)199 ± 20 ± 00 ± 042 ± 15519 ± 17364 ± 34471 ± 14n.d.n.d.
 Prolactin (µg/L)n.d.n.d.n.d.n.d.n.d.n.d.n.d.51 ± 7156 ± 50
  • LVEDD, left ventricular enddiastolic diameter; LVESD, left ventricular endsystolic diameter; IVSD, interventricular endsystolic diameter; IVDD, interventricular enddiastolic diameter; LVPWD, left ventricular posterior wall diameter; LVPS, left ventricular posterior septal diameter; FS, fractional shortening; LVEF, left ventricular ejection fraction; CK, creatine kinase; CK-M, creatine kinase muscle; ASAT, aspartate transaminase; ALAT, alanine transaminase; LDH, lactate dehydrogenase; NT-pro-BNP, N-terminal pro brain natriuretic peptide; CRP, c-reactive protein; n.d., not determined; PPCM, peripartum cardiomyopathy; PCTR, age-matched healthy pregnant; PPCTR, healthy post-partum; NPCTR, healthy non-pregnant women; ICM, ischaemic cardiomyopathy; CAD, coronary artery disease; HCTR, age-matched healthy controls. Plus-minus values are means +SD.

Isolation of microparticles

Isolation of the MPs was performed using one unique protocol in our laboratory only in order to rule out variations in pre-analytic procedures. Whole blood in EDTA (5 mL) of all samples was centrifuged at 1.500 g for 15 min to prepare platelet-rich plasma, centrifuged again for 2 min at 13.000 g to obtain platelet-poor plasma within 30 min of blood sampling, and stored at −20°C for 1 week and afterwards at −80°C until analysis. Before storage, aliquots of 125 µL platelet-free plasma and 25 µL platelet-enriched plasma sample sizes were assembled. Analysis was always performed after 28–35 days to minimize differences due to different long period of thawing time.36,37 Total 125 µL of platelet-poor plasma was partitioned into five tubes (25 µL each). On with fluorescent monoclonal antibodies (2 µL each): phycoerythrin (PE)-labelled anti-CD31 (BD Biosciences, San Jose, CA, USA), fluorescein isothiocyanate-labelled (FITC) anti-CD144 (R&D, Minneapolis, MN, USA), FITC-labelled anti-CD62E (BD Biosciences), FITC-labelled anti-CD14 (BD Biosciences), PE-labelled anti-CD45 (BD Biosciences), and allophycocyanin-labelled AnnexinV (BD Biosciences). Platelet-derived microparticles were analysed in platelet-rich plasma by incubation with FITC-labelled anti-CD62P, and allophycocyanin-labelled AnnexinV (all BD Biosciences). The samples were incubated at room temperature for 30 min with gentle shaking (orbital shaker, 1500 g). Phosphate-buffered saline buffer (0.20 µm filtered for reducing background noise23) was added to make the total volume 1 mL, and the samples were then analysed on flow cytometer (FACS Calibur, BD Biosciences) and Cell Quest Pro software to detect fluorescence, forward, and sideward scatter. For exact delineation of CD31-positive EMPs and not platelet-derived CD31-positive MPs, CD42b-negative MPs were analysed in platelet-free plasma. For definition of pure EMPs, CD144 was chosen additionally, as this marker is selectively expressed on endothelial cells.24 For distinguishing between activation and apoptosis in terms of EMP generation, CD62E and CD31 have been chosen. It is known that EMPs expressing constitutive markers, such as CD31, are markedly increased in apoptosis, whereas those expressing inducible markers, such as CD62E, are increased in activation.10 Therefore, phenotypic assessment of EMPs could provide relevant information reflecting the nature of endothelial injury. Markers of endothelial and platelet activation have been reported to be elevated in patients with congestive heart failure.1113 As some of these markers provided prognostic information independent of LV ejection fraction, we evaluated in our measurements whether one of these EMP subsets could be identified as specific biomarker for PPCM.

Flow cytometry

Microbeads from an FACS size-calibration kit (LB-30, Sigma, Munic) were used for size calibration. Logarithmic scale was implemented for forward scatter signal, side scatter signal, and each fluorescent channel. Non-stained samples and isotype controls were used to discriminate true events from background noise, and to increase the specificity for MP detection. An isotype control antibody was used as a negative control in all measurements and subtracted from MP counts. In all our measurements, non-specific binding was accounted for less than 7% of total MPs. Endothelial microparticles smaller than 1 µm were quantified in subpopulations, that is, CD31+CD144+AV+, CD31+CD144+AV, CD31+AV+, CD31+CD42bAV+, and CD62E+ MPs. Both CD31 and CD144 are considered endothelial markers but CD144 is known to be a more endothelial-specific marker as CD31 is also expressed on platelets. Therefore, double-positive MPs for CD31 and CD144 were defined as most specific EMPs. Expression of constitutive markers, such as CD31, are markedly increased in apoptosis, whereas those expressing inducible markers, such as CD62E, are increased in activation. For distinguishing between apoptotic and activated EMPs the ratio CD62E/CD31 was calculated.17 As previously described10 and since then controversially discussed, that only a small proportion of MPs in reality bound annexin V, both subsets have been analysed, CD31+CD144+AV+ and CD31+CD144+AV EMPs. Platelet-derived microparticles smaller than 1 µm were quantified in a specific population of CD42b+CD62P+AV+ MPs. Values were reported as counts in 1 µL platelet-poor or platelet-rich plasma (counts/500 000 events). MPs and calibrator beads (CAL) (10 µm diameter) were visualized in a forward light scatter (FSC) and side-angle scatter (SSC) as shown in Figure 1. MPs were defined as events (size 0.1–1 µm, R1) and then plotted in the ‘R2’ window (upper left) on an FL/FSC fluorescence dot plot to determinate positively labelled MPs by specific antibodies. Microparticles concentration was assessed by comparison to flowcount calibrator beads. Events beyond the gated MPs representing background noise due to logarithmic measurements were excluded in counts. Laboratory personnel who performed the blood assays were unaware of any subject's clinical or laboratory data.

Figure 1

(A) Representative FACS dot plots of microparticles in plasma from one patient with peripartum cardiomyopathy (B and C) and from one healthy pregnant woman (D). Microparticles were isolated from EDTA plasma, and analysed by flow cytometry as described in methods. Microparticles and calibrator beads (CAL) (red dots, each 3 µm diameter size) are visualized in a forward scatter (FSC)/side scatter (SSC) logarithmic representation in R1 (A). Microparticles are defined as events (size 0.1–1 µm) gated in the ‘R2’ window (upper left). (C and D) Representative FACS dot plot of increased CD31+AnnexinV+ EMPs in plasma from one patient with PPCM (C) in contrast to a healthy pregnant women (D).

Statistical analysis

Data are expressed as mean ± standard error of the mean (SEM). Continuous variables were tested for normal distribution with the Kolmogorov–Smirnov test and compared using a two-way ANOVA test, followed by a two-sided Bonferroni post hoc testing. Tests for equal variance normality were performed using the Levene Median test. A P-value of <0.05 was considered statistically significant. Assumptions of normality and equal variance were automatically tested using the statistic programme. Normal distribution of the parameters (NT-pro-BNP, CRP) was tested here using a Kolmogorov–Smirnov test. Both parameters showed normal distribution which is why they are reported as mean ± SD. Statistical analyses were performed using SigmaStat version 3.5. All data analyses and event classifications were performed by investigators blinded to the MP status of the patients and controls.


Baseline data

Baseline data of the 158 study patients are given in Table 1. FACS-analysis, as shown in Figure 1 with representative dot plots, could be performed in all 158 samples. Peripartum cardiomyopathy patients, PPCTR, PCTR, and NPCTR in one group and PPCM-BR and PPCM-BRCTR in another group were age-matched. Controls had normal LV fractional shortening and diameters without any differences between the groups. Patients with CAD, ICM (CAD plus chronic heart failure) as well as age-matched older controls (HCTR) were older but also age- and gender-matched within their groups. In ICM, patients impairment in ejection fraction was similar to PPCM patients.


Microparticles of all cell-origin, except for leucocyte MPs (LMPs), were found consistently and strongly elevated in PPCM patients compared with the other in this study included six groups.

Endothelial microparticles

In detail, total number EMPs (CD144+) were robustly elevated (P< 0.001) in PPCM compared with healthy post-partum, pregnant, non-pregnant, and controls (PPCTR, PCTR, NPCTR, and HPCTR) as well as compared with patients with ICM or with CAD (Figure 2A and Table 2). In pregnancy itself (PPCTR and PCTR), EMPs were increased to values similar to ICM or CAD, while the numbers of EMPs were significantly lower in PPCTR, NPCTR, and HCTR at older age (Figure 2A and Table 2). A subgroup of EMPs, double-positive for CD144 and CD31-staining showed legally a significant increase in PPCM (Table 2), but at lower levels in total. As shown in Figure 2B, activated EMPs (CD62E+) were significantly increased in PPCM (P< 0.001, Table 2). In contrast, apoptotic EMPs (CD31+AV+) were elevated in PPCM and PCTR compared with age-matched PPCTR and NPCTR, whereas they were higher in patients with ICM and CAD compared with PPCTR, NPCTR, and HCTR (P= 0.002; Figure 2C and Table 2). Apoptotic and activated EMPs were nearly the same in pregnant healthy women (Figure 2B and C and Table 2). Therefore, the increase in EMPs in PPCM is mainly due to an increase in activated EMPs, which is reflected in a higher CD62E/CD31 ratio (Table 3). In PPCM, the ratio is almost three-fold, which increases following treatment with bromocriptine. A further EMP-subpopulation (CD31+CD144+AV) revealed no alteration in all groups.

View this table:
Table 2

Cellular microparticles of patients with peripartum cardiomyopathy, post-partum, pregnant, non-pregnant healthy controls, patients with ischaemic cardiomyopathy, patients with stable coronary artery disease and healthy controls, peripartum cardiomyopathy treated with bromocriptine, peripartum cardiomyopathy not treated with bromocriptine

m ± SEMm ± SEMP-value vs. PPCMFold- increase SEMm ± SEMP-value vs. PPCMFold- increase SEMm ± SEMP-value vs. PPCMFold- increase SEMm ± SEMP-value vs. PPCMFold-increase SEM
 CD144+/AV+62 957 ± 87465184 ± 456<0.00112.1 ± 19.220 300 ± 5760<0.0013.1 ± 1.53963 ± 428<0.00115.9 ± 20.416 216 ± 1926<0.0013.8 ± 4.5
 CD144+/CD31+/AV+34 946 ± 30092589 ± 270<0.00113.5 ± 11.117 173 ± 2047<0.0012.0 ± 1.62916 ± 159<0.00112.0 ± 18.912 093 ± 1098<0.0012.9 ± 2.4
 CD31+/AV+10 455 ± 19681827 ± 148<0.0015.7 ± 10.28720 ± 8610.4371.2 ± 1.71438 ± 326<0.0017.2 ± 4.65291 ± 8530.0311.9 ± 1.7
 CD62E+31 946 ± 47332719 ± 2120.00211.8 ± 22.38840 ± 1669<0.0013.6 ± 2.81317 ± 212<0.00124.3 ± 22.310 448 ± 22340.0023.0 ± 2.0
 CD62P+/CD42b+/AV+19 861 ± 21422730 ± 320<0.0017.2 ± 6.69600 ± 1442<0.0012.1 ± 1.53372 ± 1229<0.0015.9 ± 1.79285 ± 426<0.0012.1 ± 5.0
 CD62P+14 952 ± 1515818 ± 70<0.00118.3 ± 21.66966 ± 34480.0162.1 ± 0.41974 ± 905<0.0017.6 ± 1.66405 ± 542<0.0012.3 ± 2.7
 CD42b+/AV+4942 ± 15532181 ± 7360.0032.3 ± 2.1433 ± 3120.0043.5 ± 4.91752 ± 265<0.0012.8 ± 2.73443 ± 8350.0152.8 ± 1.9
 CD14+/AV+15 321 ± 27512461 ± 236<0.0016.2 ± 11.66297 ± 1520<0.0012.4 ± 1.83121 ± 181<0.0013.3 ± 15.17668 ± 751<0.0012.0 ± 3.7
 CD45+/AV+5211 ± 24473360 ± 613<0.0011.6 ± 3.95159 ± 6440.9861.0 ± 3.84557 ± 7660.6171.1 ± 3.18528 ± 3830.0110.6 ± 6.3
m ± SEMm ± SEMP-value vs. PPCMFold- increase SEMm ± SEMP-value vs. PPCMFold- increase SEMm ± SEMP-value vs. PPCMFold- increase SEMm ± SEMP-value vs. PPCMFold- increase SEM
 CD144+/AV+62 957 ± 87467127 ± 427<0.0018.8 ± 20.71072 ± 679<0.00158.7 ± 12.95241 ± 742<0.00112.0 ± 11.863 951 ± 87690.1371.0 ± 0.9
 CD144+/CD31+/AV+34 946 ± 30097067 ± 622<0.0014.9 ± 4.81688 ± 991<0.00120.7 ± 3.04059 ± 17<0.0018.6 ± 3.039 842 ± 31430.8020.9 ± 0.9
 CD31+/AV+20 457 ± 15094775 ± 17990.0344.3 ± 0.81114 ± 674<0.00118.3 ± 2.2918 ± 86<0.00122.3 ± 17.510 088 ± 9440.8021.8 ± 0.1
 CD62E+31 946 ± 47334319 ± 16320.0017.3 ± 2.91572 ± 963<0.00120.3 ± 4.95750 ± 20840.0013.6 ± 2.328 656 ± 19330.5381.2 ± 0.3
 CD62P+/CD42b+/AV+19 861 ± 21422134 ± 480<0.0019.3 ± 4.43804 ± 775<0.0015.2 ± 2.74023 ± 1398<0.0014.9 ± 1.517 992 ± 22950.0761.1 ± 0.9
 CD62P+14 952 ± 15151177 ± 170<0.0012.7 ± 8.92866 ± 352<0.0015.2 ± 4.31880 ± 218<0.0018.0 ± 6.914 084 ± 18120.2601.1 ± 0.8
 CD42b+/AV+4942 ± 15531699 ± 1330.0032.9 ± 11.64440 ± 8970.4301.1 ± 1.7843 ± 290<0.0015.9 ± 5.34946 ± 7900.9931.0 ± 1.9
 CD14+/AV+15 321 ± 27513183 ± 759<0.0014.8 ± 3.64278 ± 941<0.0013.6 ± 2.9n.d.n.d.n.d.n.d.n.d.n.d.
 CD45+/AV+5211 ± 24473017 ± 358<0.0121.7 ± 6.84062 ± 9810.2311.2 ± 2.4n.d.n.d.n.d.n.d.n.d.n.d.
  • Shown are mean cell counts per 500 000 events and SEM next to P-value of each group vs. PPCM and its fold-increase vs. PPCM.

  • MP, microparticles; EMPs, endothelial microparticles; PMP, platelet-derived microparticles; MMPs, monocyte microparticles; LMPs, leucocyte microparticles; m, mean; SEM, standard error; PPCM, peripartum cardiomyopathy; PPCTR, post-partum; PCTR, pregnant; NPCTR, non-pregnant healthy controls, ICM, ischaemic cardiomyopathy; CAD, coronary artery disease; HCTR, healthy controls; PPCM-BR, peripartum cardiomyopathy treated with bromocriptine; PPCM-BRCTR, peripartum cardiomyopathy not treated with bromocriptine.

View this table:
Table 3

CD62/CD31 ratio in endothelial microparticles in patients with peripartal cardiomyopathy compared with age-matched post-partum, healthy pregnant and healthy non-pregnant women, patients with ischaemic cardiomyopathy, stable coronary artery disease age-matched healthy controls, peripartum cardiomyopathy treated with bromocriptine, and peripartum cardiomyopathy not treated with bromocriptine

CD62E/CD31 Ratio SEMP-value vs. PPCM
PPCM3.2 ± 0.8
PPCTR1.5 ± 0.20.018
PCTR1.0 ± 0.10.002
NPCTR0.9 ± 0.1<0.001
ICM2.0 ± 0.50.204
CAD1.1 ± 0.70.026
HCTR1.5 ± 0.50.033
PPCM-BR6.1 ± 1.00.004
PPCM-BRCTR2.9 ± 0.30.870
  • This ratio is used as an index of activation (high ratio, >4) or apoptosis (low ratio, <0.4) for distinguishing between apoptotic or activated EMP generation. In patients with PPCM (PPCM, PPCM-BR, and PPCM-BRCTR) a higher ratio of activation was measured in contrast to the other groups.

  • PPCM, peripartum cardiomyopathy; PPCTR, post-partum; PCTR, pregnant; NPCTR, non-pregnant healthy controls, ICM, ischaemic cardiomyopathy; CAD, coronary artery disease; HCTR, healthy controls; PPCM-BR, peripartum cardiomyopathy treated with bromocriptine; PPCM-BRCTR, peripartum cardiomyopathy not treated with bromocriptine.

Figure 2

Circulating microparticles (MP) in patients with peripartal cardiomyopathy (PPCM) compared with age-matched post-partum (PPCTR), healthy pregnant (PCTR) and healthy non-pregnant women, patients with ICM, stable CAD, and age-matched healthy controls (HCTR). (A) Endothelial microparticles (EMPs) were significantly increased in PPCM in contrast to PPCTR, PCTR, NPCTR, and ICM (P< 0.001) and patients with stable CAD (P< 0.003). Endothelial microparticle levels were not different in ICM and CAD, but significantly elevated in contrast to healthy age-matched controls without CAD (P< 0.004). (B and C) CD62E-positive EMPs represent activated EMPs, whereas CD31/AV-positive EMPs imply apoptotic EMPs. Activated EMPS were significantly higher detectable than apoptotic EMPs in PPCM than in all other groups (P< 0.001). In pregnant healthy controls, apoptotic and activated EMPs were measurable to the same extent. (D) PMPs were significantly increased in patients with PPCM in contrast to PCTR, NPCTR, ICM, CAD, and HCTR (P< 0.001). (E and F) CD62P-positive PMPs represent activated PMPs whereas CD42b-AV-positive PMPs imply apoptotic PMPs. Activated PMPs were significantly higher detectable than apoptotic PMPs in PPCM than in pregnant (P< 0.005) or non-pregnant healthy controls (P< 0.001). In PCTR significantly higher levels of activated PMPs could be identified than in PPCTR and NPCTR controls whereas measurements of apoptotic PMPs showed the same levels in these three groups. (G) MMPs were significantly increased in patients with PPCM in comparison with PPCTR, PCTR and NPCTR (P< 0.001) whereas LMPs (H) show no alteration among PPCM, PCTR, PPCTR, CAD and HCTR but a significant nearly two-fold increase in ICM (*PPCM vs. PTCR, #PPCM vs. NPCTR, §PCTR vs. NPCTR, $PPCM vs. ICM, &PPCM vs. CAD, %PCM vs. HCTR, ICM vs. HCTR, PCTR vs. HCTR, CAD vs. HCTR, ¢ICM vs. PCTR, ΩICM vs. NPCTR, ¥CAD vs. PCTR, ΨCAD vs. NPCTR, £ICM vs. PPCM, PCTR vs. CAD, ICM vs. CAD, HCTR vs. PCTR, PPCM vs. PPCTR, «PCTR vs. PPCTR, æICM vs. PPCTR, @CAD vs. PPCTR).

Platelet-derived microparticles

Platelet-derived microparticles (CD62P+CD42b+AnnexinV+) were detectable at significantly higher levels in PPCM than in healthy post-partum, pregnant, non-pregnant, and controls (PPCTR, PCTR, NPCTR, and HCTR) and in patients with CAD (P< 0.001; Figure 2D and Table 2).

Activated PMPs were significantly increased in PPCM compared with PPCTR (P< 0.001), to PCTR (P< 0.001), to NPCTR (P< 0.001), to ICM (P< 0.001), and to CAD (P< 0.001, all shown in Figure 2E and Table 2). Apoptotic PMPs were significantly altered in comparison with PCTR (P= 0.004), with PPCTR (P= 0.003), with NPCTR (P< 0.001), ICM (P= 0.015), and CAD (P= 0.003), as demonstrated in Figure 2F and in Table 2. Apoptotic PMPs in PPCM were not significantly increased compared with HCTR (P= 0.430; Figure 2F and Table 2).

Monocyte microparticles and leucocyte microparticles

Monocyte microparticles (CD14+AnnexinV+) were significantly increased in PPCM compared with all other groups (P< 0.001; Figure 2G and Table 2). Leucocyte microparticles, CD45+AnnexinV+, were significantly increased compared with PPCTR (P< 0.001) but not significantly augmented compared with the other groups. In contrast, they were increased in ICM compared with CAD (P= 0.012) and HCTRs (P= 0.003; Figure 2H).

Endothelial microparticles and platelet-derived microparticles after ablaction with bromocriptine

Peripartum cardiomyopathy patients treated with bromocriptine showed significantly improved LV ejection fraction and heart failure symptoms.7 Importantly, in these patients, EMPs were significantly decreased compared with PPCM patients with standard heart failure therapy only (P< 0.001; Figure 3 and Table 2). Beyond this, PMPs were significantly decreased in patients treated with bromocriptine (P< 0.001; Figure 3 and Table 2) in contrast to patients with PPCM and only receiving heart failure therapy.

Figure 3

Circulating microparticles (MP) in patients with peripartal cardiomyopathy (PPCM) treated with bromocriptine (PPCM-BR) compared with PPCM patients with only heart failure therapy (PPCM-BRCTR). Endothelial microparticles in patients treated with bromocriptine were significantly decreased compared with PPCM patients only treated with heart failure therapy (P> 0.001). Among these, as well CD62E-positive (activated) as CD31-positive (Apoptotic) EMPs showed significantly lower levels in PPCM-BR. In PPCM-BR significantly lower levels of PMPs in total and activated or apoptotic PMPs could be identified than in PPCM-BRCTR.


This study demonstrates for the first time that endothelial and platelet MPs isolated from peripheral blood are specifically elevated in patients with PPCM compared with pregnant, post-partum, and non-pregnant female controls next to patients with known vascular disease such as ICM and stable CAD. Microparticle expression profiles differentiated PPCM patients from cardiomyopathies of cardiovascular causes such as ICM. Therapeutic intervention using bromocriptine decreased levels of EMPs in these patients.

Peripartum cardiomyopathy is a rare and life threatening cardiomyopathy affecting young women during pregnancy or early in puerperium.37,2529 Associations with myocarditis30 or rare mutations of familiar cardiomyopathies31 were reported and genetic backgrounds appeared to be involved.32,33 Recently, it was discovered that an antiangiogenic and proapoptotic 16 kDa cleavage protein from the nursing hormone PRL is involved in the specific pathophysiological mechanism of PPCM.4,6 This discovery was made in mice with a cardiac-specific deletion of the signal transducer and activator of transcription-3 (STAT 3) which triggers the activation of CD, an ubiquitous lysosomal enzyme that cleaves 32 kDa PRL into its toxic 16 kDa form producing endothelial damage and inflammation with subsequent myocardial dysfunction and heart failure. The pathomechanism of PPCM is centred on endothelial cell damage, inflammation, and oxidative stress.4,6 Markers associated with these processes like apoptotic and activated EMPs as well as PMPs and to a lesser extent MMPs were elevated in the present study. These findings mirror the pathophysiology of endothelial cell damage followed by platelet activation and inflammation. The observed MP pattern in peripheral blood clearly dissociates PPCM from normal pregnancy, CAD and also CAD with heart failure (ICM). Furthermore, there were no differences in young non-pregnant women and elderly non-diseased individuals indicating that age does not significantly contribute to this finding. Therefore, it is shown that the MP profile, in particular EMPs and PMPs, indicate specific markers for PPCM.

With the discovery of the pathomechanism, a therapeutic target for PPCM has been identified. Inhibition of PRL secretion by bromocriptine has been shown to reduce cardiovascular events like death,34 severe heart failure and the decline of ejection fraction.6,7,34,35 Therefore, a marker enabling early detection of PPCM and to distinguish it from other changes of pregnancy like fluid overload with peripheral oedema, shortness of breath and alteration of end-diastolic volumes is needed for early diagnosis and treatment of this devastating condition, which is a major problem both in the US and in developing countries such as South Africa or Haiti.36,37 Previously, markers have been shown to be activated in PPCM.38 However, markers like brain natriuretic peptides, Apo-1, Fas-Apo-1, TNFα, and CRP are not able to distinguish PPCM from pre-existing cardiomyopathy, which is not responsive to inhibition of PRL secretion with bromocriptine.6,7,38 An early specific diagnosis of PCCM is essential for the initiation of life-saving treatments like inhibition of PRL with bromocriptine, whose effect is also mirrored by the MP profile. Therefore, cellular markers which specifically reflect the pathophysiological events involved in PPCM as reported here providing a pattern which might be helpful to early diagnose this condition and to pave the way for early treatment.

We were able to study patients in whom bromocriptine was used to block the nursing hormone PRL. Endothelial microparticles were significantly decreased in PPCM treated with bromocriptine compared with patients with PPCM with only heart failure therapy. This pronounced decline of EMPs adds to the pathophysiological concept of endothelial damage by PRL-cleavage sensitive to the treatment with bromocriptine in the pathomechanism of PPCM.

Even though these data are very promising, one limitation has to be taken into account concerning EMP measurements before and after bromocriptine treatment. An intra-individual comparison of activated and apoptotic EMPs before and after treatment would certainly strengthen the data. Unfortunately, the samples available for this very first report of an effect of bromocriptine on MP composition in patients with PPCM did not provide this. Longitudinal measurements will have to be further investigated.

A further limitation of our study concerns the measurements from frozen samples. Indeed, many groups have investigated the influence of methodological aspects such as centrifugation, freezing temperature21,22 and time next to blood sampling and the use of buffers and isotypes.23,39 Within these known limits, our measurements have all been performed according to one protocol in our laboratory in terms of thawing procedure and definite time of analysis. A pre-requisite for our analyses is the stability of MPs after a freezing procedure. Tests prior to the study showed that MPs stay stable during long-term freezing between 26 and 39 days. Our findings correspond to methodological findings of other groups.22,23,39

Another limitation of the present study is that a specific role of MPs is also possible and has not been characterized in depth. Microparticle formation and shedding involves changes of the structural architecture with disruption of the cytoskeleton after cell activation and apoptosis in which shear stress40 and proinflammatory cytokines may play a role.41,42 Microparticles contain cytokines producing biological effects like oxidative stress42 or vasoconstriction.43 This study does not allow conclusions on the potential role of MP changes predicting prognosis. However, the ratio of endothelial progenitor cells and MPs was able to predict prognosis in vascular patients.44,45 Here, more prospective studies are needed.

In conclusion, our data provide for the first time evidence that endothelial and platelet MP changes specifically indicate PPCM and might facilitate the diagnosis of this life threatening condition thereby clearly distinguishing it from heart failure of other causes and vascular disease in the absence of heart failure or pregnancy. An early diagnosis of PCCM is essential for the initiation of life-saving treatments like inhibition of PRL with bromocriptine, whose effect is also mirrored by the MP profile.

Finally, PPCM is known to be the most common cardiovascular cause of severe complications in pregnancy. The recently published guidelines46 have emphasized the great relevance of such guidelines on disease management in pregnancy and gave special consideration to the fact that all measures concern not only the mother, but the foetus as well. The optimum treatment of both must be targeted. Diagnostic procedures, interventions, and therapies should be performed according to the guidelines in order to save these young women and their newborn.

Identification of a specific MP profile for PPCM will give the possibility not only to identify patients with PPCM as early as possible and to initiate the life-saving therapy, but also to distinguish PPCM from other cardiomyopathies. This could become important in terms being a tool not only for identification, diagnostic, and differentiation but moreover for prognostic value in heart failure diseases.


K.W. and M.B. were supported by the Ministry of Science and Economy of the Federal State of the Saarland and the HOMFOR-program. K.W. and M.B. are supported by the Federal Ministry of Education and Research (BMBF). M.B. is supported by the Deutsche Forschungsgemeinschaft (KFO 196). K.S. and D.H.-K. are supported by a grant from the South African National Research Foundation and Deutsche Forschungsgemeinschaft on PPCM.

Conflict of interest: none declared.


View Abstract