This prospective study included 85 patients with cardiogenic shock hospitalized at the Coronary-Care Unit (CCU) of the University Hospital in Pilsen, Czech Republic between March 2017 and April 2018. We used following criteria for cardiogenic shock: (i) systolic blood pressure < 90 mmHg for > 30 min or vasopressors required to achieve a blood pressure ≥ 90 mmHg; (ii) pulmonary congestion or elevated left-ventricular filling pressures; (iii) signs of impaired organ perfusion with at least one of the following criteria: (a) altered mental status; (b) cold, clammy skin; (c) oliguria; (d) increased serum-lactate. We included patients in whom invasive blood pressure monitoring was indicated. None of the patients received an intra-arterial catheter only for the purpose of the study. The management of the patient was guided exclusively by invBP with the target invasive mean arterial pressure (invMAP) value ≥70 mmHg. Exclusion criteria were as follows: extreme obesity (body mass index > 40 kg/m2), use of extracorporeal membrane oxygenation, intra-aortic balloon counter-pulsation, and cannula insertion into other than the left or right radial artery.
Blood pressure measurement
We started BP recordings as soon as the patient’s condition was stabilized (in all patients within three hours of the initiation of medical care).
Order of BP measurements
First was the invasive BP value; next were two auscultatory BP measurements; followed by two oscillometric measurements; and lastly the second invasive BP measurement was recorded. For BP comparisons, the mean values of each pair of BP measurements were used. BP measurements were repeated in the same patient every six hours for 72 h; thus we expected to obtain12 sets of BP measurements for each patient (1020 sets of measurement). However, seven patients were dismissed from the CCU before end of the study. Moreover, in four cases we were not able to measure oscillometric BP. The values of invasive and auscultatory BP in these cases were: 108/79 and 110/70; 50/20 and 40/20; 105/78 and 110/70; 102/65 and 100/60 mmHg. In two times atrial fibrillation was present, in other two cases frequent supraventricular of ventricular extrasystoles were present. Therefore, number of BP sets totaled 967. In non-invasive measurements, systolic (SBP) and diastolic BP (DBP) were recorded and mean arterial pressure (MAP) calculated according to the formula: MAP = DBP + ((SBP-DBP) ÷ 3). For invasive measurements, we directly recorded MAP, SBP, and DBP. Only nine experienced nurses (i.e., shift supervisors) were responsible for taking all BP measurements. At the time of each BP measurement, the following were are also recorded: the patient’s Glasgow coma score, body temperature, electrocardiogram, Richmond Agitation-Sedation Scale (RASS), information on patient sedation, and whether the patient had received (was receiving) vasopressive or inotropic medications.
Invasive blood pressure measurement
A 20-gauge cannula was inserted into the right or left radial artery and connected to a disposable pressure transducer (Combitrans Monitoring Set (arterial); B. Braun Melsungen AG, Germany) using 100-cm-long tubing. The transducer system was set up by an experienced nurse and checked by a physician in all cases. The forearm was at the same level as the brachial cuff (i.e., at the level of the phlebostatic axis) to eliminate hydrostatic pressure. Air bubbles were carefully flushed from the system before data collection. The zero level for the arterial blood pressure was taken at the right atrium (i.e., at the level of the phlebostatic axis) and the arterial waveform was displayed on a Solar 8000i monitor (GE Medical Systems Information Technologies Inc., WI, United States). The average resonance frequency of the catheter tubing-transducer system was 20 Hz (range, 16–25 Hz), combined with a damping coefficient of 0.3. In addition, the correct shape of the BP waveform was ascertained using the “fast flush” test. BP values were measured at the end of expiration assessed using capnography waveform and with the help of monitor freezing at the time of BP reading to minimize the effect of changing intrathoracic pressure on invasively measured BP values . In patients with regular heart rhythm we recorded one invasive BP value just before and just after non-invasive BP was taken. In those cases when arrhythmia was present, we recorded three consecutive invasive BP values at the beginning and three consecutive invasive BP values at the end of BP measurement; from these values we calculated mean used in analysis.
Non-invasive blood pressure measurement
A dual tonometer Nissei DM-3000 (NISSEI, Nihon Seimitsu Sokki Co., Japan) was used for taking non-invasive blood pressure measurements . According to the manufacture’s recommendations, the patient’s mid-arm circumference was used to determine the appropriate brachial cuff size. The brachial cuff was placed on the same arm as the radial artery catheter. Two BP auscultatory measurements, followed by two oscillometric measurements were taken, with 1-min intervals between each measurement.
Baseline biochemical variables (blood samples were drawn at the time of the first BP measurement) were measured. NT-pro BNP values were determined using original analytical kits from Roche on a Cobas 8000 analyzer. High-sensitivity cardiac troponin I (hsTnI) was measured using the Architect i2000 platform and a STAT High Sensitive Troponin-I assay (Abbott Diagnostics, USA). We measured serum lactate and blood pH initially, and 12, 24, and 48 h after admission to the CCU.
Left ventricle ejection fractions were recorded using bed-side echocardiograms, which were acquired using a Vivid 7 ultrasound system (GE Medical Systems, Horton, Norway) with a 3.4-MHz multi-frequency transducer. Left ventricular filling pressure was also estimated using echocardiography. We evaluated peak left ventricular filling velocity during early diastole (E wave), during atrial contraction (A wave) and deceleration time (DT). DT value < 150 ms in patients with decreased left ventricle (LV) ejection fraction was considered as a high filling pressure LV (> 25 mmHg). Furthermore, we evaluated the motion of the mitral anulus at its septal and lateral margins using pulse Doppler tissue echocardiography. As a normal diastolic velocity e‘of the lateral margin we considered ≥10 cm/s and septal ≥7 cm/s, the reduction of velocity e‘was one of the criteria of increased LV filling pressure. We calculated the E/e ‘ratio. The value of the E/e’ ratio (e’ as the average velocity of the two annular edges) ≤ 8 was indicative of the low LV filling pressure, the value > 14 reflected an increased pressure in the left atrium. If we evaluated only the rate of the lateral edge of the mitral anulus, the value > 13 was considered to be a sign of increased LV filling pressure. For the calculation using the septal margin only, the value of > 15 was considered as increased one. We considered 8–13 as a gray zone and we did not take it as a sign of increased filling pressure LV.
Sample size calculation for Bland-Altman analyses was based on estimated mean difference 5 mmHg with SD of 10 mmHg. It was estimated that 83 pairs of independent measurements would be required for 2-tailed α of 0.5 and a 1-β of 80%. Bland-Altman analyses were performed using MedCalc Statistical Software version 18.11.3 (MedCalc Software bvba, Ostend, Belgium). SAS software version 9.4 (SAS Institute Inc., USA) was used for data management and other statistical analyses. Results are presented as the arithmetic mean ± standard deviation, median with inter-quartile range (IQR) or as a proportion (percentage); differences among groups were assessed using the paired Student’s t test, the Kruskal-Wallis test, and the χ2 test, respectively. We used the two-tailed test for P value calculations.
Agreement between two different methods of BP measurement corrected for repeated measurements in one subject, using Bland-Altman analysis  and concordance correlation coefficients , was tested. Percentage error was calculated as: (non-invasive BP – invasive BP)/invasive BP * 100.The accuracy of BP measurements was also estimated according to the British Hypertension Society (i.e., a minimum percentage of readings must be within 5, 10, and 15 mmHg. For example, to meet grade A, the absolute difference between the two methods must be less than 5 mmHg in at least 60% of measurements, less than 10 mmHg in at least 85% of measurements and less than 15 mmHg in at least 95% of measurements. To achieve grade C, the corresponding percentages are 40, 65 and 85. To achieve a grade all three percentages must be equal to or greater than the tabulated values. . We also used linear regression analysis after accounting for repeated measurement in one subject to study determinants of BP differences.