Microcirculation in women with severe pre-eclampsia and HELLP syndrome: a case–control study




To compare microcirculatory perfusion in women with severe pre-eclampsia against that in healthy pregnant women, and secondly in women with severe pre-eclampsia with or without HELLP syndrome (haemolysis, elevated liver enzymes, and low platelets).


Case–control study.


University Hospital Rotterdam, the Netherlands.


Twenty-three women with severe pre-eclampsia and 23 healthy pregnant controls, matched for maternal and gestational age. Out of the 23 women with severe pre-eclampsia, ten presented with HELLP syndrome.


Microcirculation was analysed sublingually by a non-invasive sidestream dark-field imaging device (SDF).

Main outcome measures

Perfused vessel density (PVD), microcirculatory flow index (MFI), and heterogeneity index (HI) were calculated for both small vessels (∅ < 20 μm; capillaries) and non-small vessels (∅ > 20 μm; venules and arterioles).


There were no significant differences between women with severe pre-eclampsia and healthy controls. Women with pre-eclampsia and HELLP syndrome showed a reduced PVD (P = 0.045), MFI (P = 0.008), and increased HI (P = 0.002) for small vessels, as compared with women with pre-eclampsia but without HELLP syndrome.


Sidestream dark-field is a novel, promising technique in obstetrics that permits the non-invasive evaluation of microcirculation. We did not observe major differences in sublingual microcirculatory perfusion between women with severe pre-eclampsia and healthy pregnant controls. In women with severe pre-eclampsia, the presence of HELLP syndrome is characterised by impaired capillary perfusion.


The microcirculation is a vast network of small vessels with a diameter below 100 μm. It consists of arterioles that regulate flow to the capillaries, which subsequently drain in venules.[1] Exchange of oxygen and nutrients occurs at the level of the capillaries, which mainly consist of a thin layer of endothelium. With the availability of new imaging modalities, the importance of microcirculatory perfusion in the pathophysiology, prognosis, and treatment of conditions with profound haemodynamic imbalance, like sepsis, shock, and cardiac disease, is emerging. Parameters of microcirculatory perfusion seem independent of global haemodynamic status and appear to be strong predictors of outcome.[1-6] Sidestream dark-field (SDF) imaging is a novel technique enabling direct, non-invasive visualisation of microcirculatory perfusion at the bedside in adults, children, and newborns.[1-3] Severe pre-eclampsia is characterised by a maternal haemodynamic instability caused by generalised endothelial dysfunction.[7] Many of its symptoms and complications strongly suggest microcirculatory dysfunction. A recent study indicates that capillary rarefaction precedes the clinical onset of pre-eclampsia.[8] HELLP syndrome (haemolysis, elevated liver enzymes, and low platelets) is considered an expression of disease severity.[9, 10] Although its exact pathophysiology is not completely understood, the haemolysis, platelet consumption, and liver cell necrosis might reflect a more profound disturbance in microcirculatory function. Our aim was to explore the potential and reliability of SDF in pregnant women, and to analyse microcirculatory perfusion in women with severe pre-eclampsia as compared with that in healthy pregnant women. Secondly, we investigated the influence of HELLP syndrome on microcirculation in women with severe pre-eclampsia.


Study setting

The study was conducted from November 2009 to September 2012 at the department of Obstetrics and Prenatal Medicine of the Erasmus Medical Centre of the University of Rotterdam. Twenty-three Women with severe pre-eclampsia were included. In ten of these women, pre-eclampsia was complicated by HELLP syndrome. Four women with severe pre-eclampsia had a history of systemic lupus erythematosus or chronic hypertension.

Twenty-three healthy pregnant women, matched for maternal and gestational age, were included as controls. Informed consent was obtained from all women and the study protocol was approved by the local medical ethical committee.

Severe pre-eclampsia was defined as pre-eclampsia (hypertension and significant proteinuria) with severe hypertension, and/or with symptoms, and/or with biochemical and/or haematological impairment.[11] HELLP syndrome was defined as the presence of at least two components of either haemolysis (lactate dehydrogenase, LDH ≥ 600 U/l), elevated liver enzymes (aspartate aminotransferase, AST ≥ 70 U/l), or thrombocytopenia (thrombocytes < 100 × 109/l).[10] All women with severe pre-eclampsia were categorised into severe pre-eclampsia either with or without HELLP syndrome according to previous definitions after expert agreement by three of the authors (J.C., E.H., and J.D.). Women with severe pre-eclampsia were managed according to our local protocol, as described in Appendix S1.

In women with severe pre-eclampsia we aimed to perform microcirculatory analysis at time points when disease activity was estimated to be maximal and when the interference from treatment was estimated to be as minimal as possible. Therefore, measurements were performed either before intravenous nicardipine, magnesium sulphate bolus, or when laboratory abnormalities consistent with HELLP syndrome occurred, irrespective of other concomitant medication. Age, parity, body mass index (BMI), gestational age, and medical history were obtained for all women.

All measurements were performed in a 15° left lateral tilt. Women were asked to refrain from eating or drinking for 30 minutes before measurements. Blood pressure was determined by manual sphygmomanometry. LDH, AST analysis, and thrombocyte count, as well as haematocrit and haemoglobin counts, were performed as part of the routine clinical procedure in women with severe pre-eclampsia on the day of the measurements.

Sidestream dark-field imaging

The sublingual microcirculation was visualized using SDF (Figure 1A).[12] This hand-held video microscope (MicroScan; MicroVision Medical, Amsterdam, the Netherlands) emits stroboscopic green light (530 nm) from an outer ring of light-emitting diodes (LEDs), which penetrates the tissue to a depth of approximately 3 mm. The light is absorbed by the haemoglobin of individual red blood cells in superficial vessels. A negative image is transmitted back, after 5× optical magnification, to an isolated synchronised charge-coupled device camera in the core of the probe.

Figure 1.

(A) Set-up for a sublingual microcirculatory perfusion measurement with SDF. A disposable sterile plastic cap covers the mouthpiece of the probe. (B) A frozen video-clip image of sublingual microcirculation as viewed with SDF.

This allows high-contrast video images of circulating erythrocytes to be recorded with a 286× magnification from the microcirculation of organs covered with a thin epithelial layer. (Figure 1B, Video S1).[5] SDF imaging has been validated against and found to be superior to intravital videomicroscopy.[1, 12]

The consensus recommendations on how to best obtain and evaluate SDF measurements were followed.[4, 13] After obtaining good image focus and contrast, with specific attention paid to avoiding pressure artefacts by assuring continuous venous perfusion, one investigator (E.H.) recorded three high-quality video clips per measurement, with a duration of at least 20 seconds, each at a different sublingual site (using a high-definition videocassette recorder: GV-HD700; Sony Instruments, Tokyo, Japan).These were digitalised, blinded, and stored on an external hard drive. After completion of the data set, E.H. performed analysis of the blinded recordings using ava 3.0 (Automated Vascular Analysis, MicroVision Medical, Amsterdam, the Netherlands).

Inter-observer variability was assessed though separate analysis of the 45 recordings of 15 randomly selected cases by a different investigator (E.B.).

As described in the consensus recommendations, the perfused vessel density (PVD), microcirculatory flow index (MFI), and the heterogeneity index (HI) for MFI were calculated, each reflecting distinctive characteristics of microcirculatory perfusion.[4, 13] Each parameter was determined separately for both small vessels (∅ < 20 μm, capillaries) and non-small vessels (20 μm ≤ ∅ ≤ 100 μm, mostly venules and arterioles).[13-15] A detailed description of these parameters and respective methods of calculation is available in Appendix S2.

Statistical analysis

Statistical analysis was performed with spss 20.0 (SPSS Inc., Chicago, IL, USA). Variables were tested for normality and compared with the Students' t-test or non-parametric Mann–Whitney U-test, as appropriate. The effect of parameters with a known potential to influence haemodynamics (gestational age, use of oral antihypertensive medication), vascular structure (maternal age, BMI, race), or SDF measurements (haematocrit, haemoglobin) was assessed by analysis of covariance (ancova) or by its non-parametric variant (the Quade test), as appropriate.[16-20] The adjusted P values, with P ≤ 0.05 (two-sided) as the limit of significance, were used without correction for multiple comparisons.

Inter-observer reliability was assessed by calculation of the intraclass correlation coefficients from each parameter (PVD, MFI, and HI), separated for small- and non-small vessels in 15 cases. Inter-observer agreement for PVD was shown in Bland–Altman plots.

In the absence of SDF data on microcirculatory perfusion in pregnancy and severe pre-eclampsia, no power calculation was performed and this study was undertaken as an exploratory pilot.


Adequate recordings and measurements were obtained for all participants. Intraclass correlation coefficients were good for capillary measurements and were moderate for larger vessels (Table 1). Figure 2 shows the inter-observer agreement for PVD in small and non-small vessels.

Table 1. Intraclass correlation coefficients (ICCs) and 95% CIs for inter-observer reliability
Small vessels0.87 (0.61–0.96)0.94 (0.77–0.98)0.96 (0.84–0.99)
Non-small vessels0.66 (0.05–0.89)0.88 (0.63–0.96)0.73 (0.15–0.91)
Figure 2.

Bland–Altman plots showing interobserver agreement for PVD in small and non-small vessels.

Twelve women with severe pre-eclampsia received concomitant oral antihypertensive medication. This included women with and without HELLP syndrome. All received methyldopa, nifedipine, or a combination of both. One woman received additional oral labetalol.

The baseline characteristics of women with severe pre-eclampsia and healthy controls were similar (Table 2). As expected, blood pressure was significantly higher in women with severe pre-eclampsia. Pre-eclampsia was considered to be severe in all women either because of the severity of their hypertension (systolic blood pressure, SBP, ≥ 160 mmHg and/or diastolic blood pressure, DBP, ≥ 110 mmHg) or because of the presence of HELLP.

Table 2. Population characteristics
CharacteristicsWomen with severe pre-eclampsia (n = 23)Controls (n = 23) P
Severe pre-eclampsia versus control pregnancies
Gestational age (weeks)a33 (21–37)33 (20–38)nt
Age (years)31 (±5)31 (±5)nt
BMIa28 (20–56)26 (18–41)ns
Systolic blood pressure (mmHg)a170 (130–2015)110 (99–135)<0.001
Diastolic blood pressure (mmHg)a102 (76–115)68 (50–90)<0.001
CharacteristicsWith HELLP (n = 10)Without HELLP (n = 13) P
  1. ns, not significant; nt, not tested (matching criterion).

  2. Values are expressed as means ± standard deviations or medians with ranges according to normality.

  3. a

    Non-parametric test used.

Severe pre-eclampsia with or without HELLP syndrome
Gestational age (weeks)a30 (21–37)33 (25–37)ns
Age (years)31.1 (±3.9)31.7 (±6.5)ns
BMIa26 (20–32)30 (20–56)ns
Systolic blood pressure (mmHg)157 (±31)174 (±20)ns
Diastolic blood pressure (mmHg)96 (±12)102 (±7)ns
Oral antihypertensive medication40%69%ns
LDH (U/l)a850 (602–2964)426 (265–575)<0.001
AST (U/l)a221 (42–1593)23 (14–52)<0.001
Thrombocytes (109/l)a99 (41–289)250 (134–375)<0.001

We could not observe any significant differences in sublingual microcirculatory perfusion in women with severe pre-eclampsia, as compared with healthy controls (Figure 3A, B).

Figure 3.

(A) Scatter plots depicting differences in perfused vessel density (PVD), microcirculatory flow index (MFI), and heterogeneity index (HI) for small and non-small vessels between women with severe pre-eclampsia and controls, as well as between women with severe pre-eclampsia with and without HELLP syndrome. *Non-parametric test used. Adjusted P values are depicted for comparisons with statistical significant difference; HELLP, severe pre-eclampsia with HELLP syndrome; PE, severe pre-eclampsia without HELLP syndrome. (B) Box plots depicting differences in PVD, MFI, and HI for small and non-small vessels between women with severe pre-eclampsia and controls, as well as between women with severe pre-eclampsia with and without HELLP syndrome. Boxes denote interquartile ranges, bars in boxes represent median values, and error bars represent ranges. *Non-parametric test used. Adjusted P values are depicted for comparisons with statistical significant difference; HELLP, severe pre-eclampsia with HELLP syndrome; PE, severe pre-eclampsia without HELLP syndrome.

Baseline characteristics between women with severe pre-eclampsia, with or without HELLP, were also comparable, except for the components of HELLP syndrome (Table 2).

Women with HELLP syndrome had significantly lower values of PVD and MFI and significantly higher values of HI for small vessels, as compared with women with severe pre-eclampsia without HELLP (Figure 3A, B). These differences remained significant after adjusting for haemoglobin count, haematocrit, BMI, medication use, pre-existent disease, maternal age, and gestational age.


Main findings

In this study we explored microcirculatory perfusion in women with severe pre-eclampsia with SDF, a novel technique in obstetrics.

Microcirculatory research has mainly been hampered by technological limitations. SDF allows the direct recording of high-contrast images and assessment of different aspects of microcirculatory perfusion. In our study, satisfactory images were obtained at the bedside and with minimal discomfort in all women. Inter-observer variability showed good reliability for capillary vessels, but is less evident in non-small vessels given the wide confidence intervals for PVD and HI, despite acceptable intraclass correlation coefficients. These findings are in line with previous results in non-pregnant populations.[4, 15, 21, 22] Fortunately previous research and our results suggest that the capillary compartment is the main area of interest in microcirculatory perfusion. Therefore, SDF seems a preferred method for microcirculatory analysis in obstetrics.[1] Nevertheless, although image recording is relatively straightforwards, offline analysis still requires substantial human input and remains time consuming. Developments in the most recent version of the SDF camera now permit automatic image analysis, which will further improve reliability, and holds promise for bedside recording and analysis in the future.

Despite the increased blood pressure we did not observe any difference in microcirculatory parameters in women with severe pre-eclampsia, as compared with healthy pregnant controls.

Apparently, the major macrocirculatory disturbances of severe pre-eclampsia are not reflected in significant differences in sublingual microcirculatory perfusion.

Interestingly, when comparing women with severe pre-eclampsia with or without HELLP syndrome, we observed significant differences in all aspects of capillary perfusion, with a decrease in PVD and MFI and an increased HI in women with HELLP syndrome.

Interpretation and relation to other studies

Previous microcirculation studies described a decreased venular diameter and increased postcapillary (venular) resistance in women with pre-eclampsia using intravital microscopy and plethysmography.[23, 24] Although we did not specifically asses changes in vessel diameters and used different techniques in different organ systems, we did not observe major changes in large vessels, which mostly consist of venules and arterioles to a lesser extent. Hasan, using intravital capillaroscopy, reported a reduced capillary density in 11 women with pre-eclampsia, as compared with normal healthy pregnant and non-pregnant women.[25] Houben, using a similar set-up, could not confirm these findings, and Vollebregt, using orthogonal polarisation spectral imaging (OPS), did not find any changes in nail-fold capillary red blood cell velocity.[24, 26] Neither did we observe any changes at a capillary level between women with severe pre-eclampsia and healthy pregnant women. This discrepancy may be explained by the use of medication in women with pre-eclampsia, as Hasan performed the measurements before any intervention. Most women in our, Vollebregt's, and Houben's studies had already received some form of antihypertensive therapy, magnesium sulphate, or steroids for fetal lung maturation. These drugs have the potential to influence capillary perfusion.[27] In further studies, attempts should be made to perform measurements before any treatment; however, this remains difficult, as the maternal condition often does not permit treatment delay in severe pre-eclampsia.

The suggestion of impaired capillary perfusion in women with both pre-eclampsia and HELLP syndrome might explain some aspects of the pathophysiology of HELLP syndrome.[10]

The reduced PVD and MFI might be a reflection of microvascular erythrocyte fragmentation and platelet adherence to the damaged endothelial surface in narrowed capillaries.[28] The increased heterogeneity could explain the diffuse pattern of liver cell necrosis in HELLP, where fibrin microthrombi and fibrinogen deposits are often observed both in intact hepatic sinusoids and in areas with hepatocellular necrosis upon histology.[9] Heterogeneity of flow is an important characteristic of impaired microcirculation.[1, 4] With heterogeneous flow, a reduced number of capillaries are perfused. Cells close to the capillaries extract the normal quantity of oxygen, but cells too far away become hypoxic. Although the total oxygen delivery is the same, heterogeneous perfusion probably affects tissue oxygenation more than a reduced but homogenous flow.

Future research

Sublingual microcirculation is easily accessible for SDF. It is representative in sepsis, probably because of the embryological and metabolic similarities with the splanchnic mucosa.[1, 4] Even so, pre-eclampsia is a complex syndrome that groups a broad clinical spectrum with variable degrees of organ dysfunction. It is therefore questionable whether the endothelial dysfunction is always manifested equally in all vascular beds. Our results, both in women with and without HELLP, could be explained by the fact that the sublingual microcirculation may not be the most representative site in all pre-eclamptic women. SDF enables microvascular analysis in different areas (e.g. skin, conjunctiva, nail fold, vagina, cervix, etc.). Further research in obstetrics should explore microcirculatory perfusion at various sites during the haemodynamic adaptation of normal pregnancy, and explore eventual representative areas in pathological conditions.

Besides facilitating (patho) physiological research in larger populations, future improved versions with rapid bedside analysis also offer perspectives for clinical implications. As in sepsis and cardiogenic shock, microcirculatory perfusion analysis has the potential to improve outcome prediction, and assist in the selection of candidates for expectant management or monitoring of medical treatment.[6, 29]

Strengths and limitations

This is the largest population of women with pre-eclampsia investigated for microcirculatory changes in a prospective, case-controlled design. Our control group of 23 pregnant women is also one of the largest investigated populations of healthy subjects using SDF. The significant capillary differences in women with HELLP syndrome seem supported by a large effect size. Although it remains controversial whether this exploratory set-up allows for adjustment, significant differences remained, irrespective of adjustment for confounding factors. The absence of clinically relevant spread in the 95% confidence intervals of most parameters suggest that the size of our population was probably sufficient to exclude differences in sublingual microcirculation between healthy women and women with pre-eclampsia. Still, the populations remain small and this study should merely be viewed as an exploratory analysis. Our results certainly need further confirmation in a larger trial, separating women with and without HELLP syndrome, and preferably before any intervention.


Sidestream dark-field (SDF) imaging is a promising technique for the study of microcirculatory perfusion in obstetrics. Our study indicates that there are no major differences in sublingual microcirculatory perfusion between women with severe pre-eclampsia and healthy pregnant controls; however, HELLP syndrome is associated with an impairment of all aspects of capillary perfusion.

Disclosure of interests

The authors have no conflicts of interest to disclose.

Contribution to authorship

JC conceived the study, analysed the data, and wrote the article and revisions. EH and EB performed and analysed the microvascular measurements, and contributed to the data analysis, writing of the article, and revisions. JD, DT, and ES significantly contributed to the study design and setting. DR and WH contributed to the statistical analysis. All authors critically revised the article and gave approval of the version to be published.

Details of ethics approval

The study was approved by the Medical Ethical Committee of the Erasmus Medical Centre of the University of Rotterdam (MEC-2007-072), as well as the Dutch Central Commission of Human Related Research (CCMO NL16284.078.07), and was registered in the European Clinical Trials Database (EudraCT 2007-000727-17).




We thank Professor Can Ince and Dr Jasper van Bommel of the Department of Intensive Care Medicine for their valuable advice during the study and for their critical appraisal of the article.