Chapter 2: Materials and methods
C- reactive protein (CRP)
High resolution CRP was measured using a Beckman Coulter assay and Beckman DxC600 analyser (Beckman Coulter Inc., Brea, CA, U.S.A). This is a turbidimetric assay, where goat and mouse anti-CRP antibody is used to bind CRP in the human serum resulting in the formation of insoluble aggregates which increase the turbidity of the sample. The change in absorbance at a 600 nanometres is proportional to the concentration of CRP. The range of detection was 1mg/L- 250mg/l and the intra and inter assay CVs were 5% and 7.5% respectively
Soluble ST2 assay
Soluble sST2 was measured using a commercially available enzyme linked immunosorbent assay (ELISA). The Critical Diagnostics Presage™ ST2 Assay (Critical Diagnostics, 3030 Bunker Hill, Suite 115A, San Diego, USA) is a quantitative sandwich monoclonal ELISA in a 96 well plate format for measurement of ST2 in serum or plasma. Diluted plasma was used
according to the manufacturer’s instructions. Precision evaluation of the assay was performed according to the Clinical and Laboratory Standards Institute (CLSI) guideline EP5-A. Four pooled patient plasma samples were aliquoted into twenty 1.5 mL plastic tubes for each concentration level and frozen at -80°C. These samples were analysed in duplicate in one run per day for 20 days. Intra-assay and inter-assay CV was calculated with the CLSI single-run precision evaluation test. The assay had a within-run CV of 6.5% and a total CV of 9.1% at a
70 mean concentration of 16.9 ng/mL (low), a within-run CV of 3.4% and a total CV of 5.5% at a mean concentration of 33.1 ng/mL (medium-1), a within-run CV of 3.8% and a total CV of 6.3% at a mean concentration of 68.7ng/mL (medium-2) and ), a within-run CV of 2.4% and a total CV of 4.8% at a mean concentration of 159.1 ng/mL (high). The normal ranges across the general population in U.S.A have been documented recently and the 95th percentile reference interval in normal, healthy males is 8.5-49.3 ng/mL (median 23.6 ng/mL), in females is 7.1-33.5 ng/mL (median 16.2 ng/mL) and in the entire group is 1.75-34.3 (median 18.8 ng/mL) 288. The measurement of sST2 was undertaken by individuals blinded to the patient characteristic and clinical data.
Radionuclide ventriculography (RNV)
Left ventricular function was evaluated by RNV performed by operators blinded to the clinical characteristics of the patients. Planar acquisition was performed with best septal separation view (32 frames, 20% window for R-R interval). SPECT was performed from right anterior oblique (RAO) 450 to left posterior oblique (LPO) 450 (16 frames, 20% window). The planar data was processed with Hermes (FUGA) and SPECT was processed with Cedars Sinai QBS software. Favourable reverse left ventricular remodelling was pre-specified as an increase of > 5
percentage points in LVEF from baseline. Inter user variability was assessed using two
operators who were fully trained in the techniques and reported the scans independently of each other. Intra observer variability was assessed by one of the operators who re-reported the scans at a later time without access to their previous measurements (Figure 10). The coefficients of variability for inter and intra observer measurements of planar RNV LVEF were 6.8% and 5.2 %
71 respectively. The coefficients of variability for inter and intra observer measurements of planar RNV RVEF were 15.7% and 7.8 % respectively.
Figure 10: Bland-Altman plots for radionuclide ventriculography
Figure 12: Bland-Altman plots. Panels A and B show inter and intra observer variability for planar
radionuclide ventriculography (RNV) measurements of left ventricular ejection fraction (LVEF) while panels C and D are for the measurement of right ventricular ejection fraction (RVEF). The middle line represents the mean difference and the outer lines the 95% confidence intervals.
A B
72
Transthoracic Echocardiography (TTE)
Transthoracic echocardiography was performed using a Phillips Sonos 5500 echocardiograph and a multifrequency transducer (Andover, MA, USA). Standard two-dimensional parasternal long and short-axis views and apical four-, five-, and two-chamber views were performed. Cross-sectional two-dimensional guided M-mode recordings of the LV minor axis were obtained. Left ventricular ejection fraction (LVEF) and end systolic volume (ESV) were calculated by Simpson’s method. Mitral regurgitation (MR) was recorded using colour-flow Doppler Mitral regurgitation (MR) was recorded using colour-flow Doppler and continuous- wave Doppler in the apical four- and two-chamber views, and its duration, expressed as percentage of the RR interval, was calculated.
Subaortic flow velocity was obtained by pulsed-wave Doppler from the apical five-chamber view, with the sample volume placed in the LV outflow tract (LVOT), 1 cm below the aortic cusps.Stroke distance was calculated as the time integral of aortic velocity, stroke volume as the product of stroke distance and subaortic area, and cardiac output as stroke volume multiplied by heart rate.
Transmitral flow velocities were recorded from the apical four chamber view using pulsed-wave Doppler. Peak E (early diastolic) was obtained in all patients and A (atrial) velocities were obtained in all patients who were in sinus rhythm.
Aortic pre-ejection time (APET) and pulmonary pre-ejection time (PPET) were measured from the onset of the QRS complex to the onsets of aortic and pulmonary flow, respectively. With
73 CRT, the pacing artefact was taken as the onset of ventricular activation. The inter -ventricular delay (IVD) was calculated as APET minus PPET.
Echo optimisation of pacemaker settings
At one month following implant all patients underwent TTE guided optimisation according to the following protocol (appendix 2). The heart rate was set at 70 bpm and transmitral pulse wave Doppler was at the mitral valve tips was obtained at a series of AV delays. The AV delay with the maximal E/A separation was selected. Subsequently the VV delay was altered in 20 ms increments. The velocity time integral (VTI) from the LVOT (5 chamber view) 1 cm below the aortic valve leaflets was measured as a surrogate of cardiac output. The setting with the maximal VTI was selected. Indices of dyssynchrony and MR were measured as detailed previously at baseline and after optimisation.
Six minute walk testing (6MWT)
Patients were asked to walk between two markers which were 10 meters apart along a level corridor in the cardiology department of our institution. They were asked to walk at the pace they felt comfortable for 6 minutes. During the test they were not encouraged and allowed to rest if they felt so inclined. Their heart rate blood pressure and oxygen saturation by pulse oximetry was measured before and after the test. The total distance covered in 6 minutes was recorded.
Quality of life questionnaires
Quality of life was assessed by the Minnesota Living with Heart failure questionnaire (MLWHF) (University of Minnesota, Minneapolis, U.S.A) which has been validated as a tool to measure quality of life in heart failure patients. The patients were asked to fill in the questionnaire
74 independently without help. The test comprises a set of 21 questions about how HF impacts on daily activities and patients are asked to mark a numeric score between 0-5 for each question. The scores were tallied to obtain a total out of 105, with the higher scores indicating a poorer quality of life attributable to HF.
Implantation of biventricular pacemakers
All implants were left sided. Endocardial right atrial and RV leads were implanted using standard techniques and access was using left subclavian vein. Those patients that fulfilled the criteria for ICD received a dual coil defibrillator lead delivered to the RV apex. The coronary sinus (CS) was intubated using standard catheters. A balloon occlusion venogram was
performed to delineate coronary sinus anatomy before delivering an LV lead to a suitable target vein. The position of the lead and choice manufacture for the device used was left to the operator’s discretion and not pre-specified in the protocol. The device was programmed to a nominal AV delay of 120 ms with simultaneous ventricular pacing until device optimisation at one month.
Statistical analysis and power calculation
Primary Endpoint
The primary outcome of the study was to be an increase in RNV LVEF of 5 percentage points. Calculating the expected sample size was performed with the help of a statistician using data from previous cohort studies of HF patients undergoing CRT. Using SPECT data the baseline EF was calculated as 23.1 % with a standard deviation of 8 % 289. Therefore to detect a
statistically significant increase of 5 % in EF (p value of 0.05) with a statistical power of 0.9 the calculated sample size is 23. Using the baseline characteristics of the MUSTIC and MIRACLE
75 trials the baseline EF was 24.4 % and 25 % respectively with standard deviations of 6.6% and 7% respectively using TTE. Therefore to detect a statistically significant increase of 5 % in LVEF (p value of 0.05) with a statistical power of 0.9 the calculated sample size is between 16 and 18.
Adjustment of sample size for calculation for neurohormones and cytokines
The group was dichotomised into responders and non-responders. As the concentrations of neurohormones were not normally distributed median apelin concentrations at the end of study was compared in the two groups by a Mann Whitney U test or log transformed data and
compared using student t tests. The same univariate statistical methods were used to compare the response to CRT with the other neurohormones, cytokines and clinical parameters which are not normally distributed. Normally distributed variables and their effect on predicting response will be evaluated by a Student's t-test. The baseline concentrations of the neurohormones were examined as to their accuracy at predicting a positive response to CRT. This was by receiver operator curve (ROC) analysis. The clinical variables measured: arterial blood pressure at implant, dyssynchrony by TTE, age, co−morbidities, baseline concentrations of neurohormones and cytokines were entered into a multivariate logistical regression analysis to determine the independent predictors of a favourable response to CRT.
It was estimated that 10% of eligible patients may fail to have a successful LV lead and in order to adjust for this fact and to enable analysis of secondary endpoints with sufficient power, we planned to recruit a sample of 100 patients.
76 The analysis of changes in clinical parameters 6 months and 12 months was performed using
ANOVA analysis of variance for normally distributed variables, with Bonferroni corrections. For variables which are not normally distributed Friedman test was used.
Patient recruitment
Twenty four patients were recruited for the study. The baseline characteristics of the group are detailed in table 5. Two patients were excluded; one patient declined CRT prior to admission and another patient had a new diagnosis of renal cell carcinoma discovered in the course of his preliminary investigations. Twenty two patients had attempted insertion of a CRT device. In three patients it was not possible site an LV lead, either due to unsuitable suitable coronary sinus anatomy or inadequate pacing parameters. In two of these patients the operators elected to place an extra lead in the RV outflow tract, to maximise separation to the existing RV apical lead and allow for multisite pacing. The other patient had an existing RV lead placed in the mid septum, and further attempts at coronary sinus pacing were abandoned. This resulted in 21 patients with multisite pacing who included in the cohort. One patient died between 6 months and 12 months and so 20 patients were available for 12 month follow up (Figure 11).
77
Table 5: Demographics of the patients recruited from the heart failure-pacing clinic
Age (years) median, (interquartile range) 74 (56-79) Male, n (%) 21 (88%) BMI mean, (SE) 28 (1.3) Heart rate (bpm) mean, (SE) 70 (2.5) Systolic BP (mmHg) mean, (SE) 117 (5.2) Diastolic BP (mmHg) mean, (SE) 68 (3.0) Mean BP (mmHg) mean, (SE) 88 (3.5) Diabetes, n (%) 4 (17%) Atrial Fibrillation, n (%) 8 (33%) Chronic right ventricular pacing, n (%) 11 (46%) QRS width (ms), mean (SE) 182 (4.3) 6 minute walk distance (m), mean (SE) 245 (69) MLWHF score, mean (SE) 48 (15)
Aetiology of heart failure n, (%)
Ischaemic heart disease 10 (42%) Dilated cardiomyopathy 11 (46%) Valvular heart disease 3 (12%) Non-ischaemic dilated cardiomyopathy (NIDCM) 14 (58%) Ventricular function calculated by radionuclide ventriculography
Left ventricular ejection fraction (%), mean (SE) 23 (1.7) Right ventricular ejection fraction (%), mean (SE) 33 (3.5) Left ventricular end diastolic volume (ml), mean (SE) 279 (27) Left ventricular end systolic volume (ml), mean (SE) 208 (24) Standard deviation left ventricular phase (0), mean (SE) 82 (5.2) Aortic pre-ejection time (ms), mean (SE) (TTE) 171 (7.0) Intra-ventricular mechanical delay (ms), mean (SE) (TTE) 53 (6.1)
78 SE=standard error, MLWHF=Minnesota living with heart failure, *the results of the patient who declined CRT are not available
Medications ACE-I / ARB n, (%) 23 (100%) Beta-blockers n, (%) 17 (74%) Aldosterone antagonists n, (%) 17 (71%) Loop diuretics n, (%) 17 (71%) Digoxin n, (%) 6 (25%) Laboratory investigations
Haemoglobin (g/dl) mean (SE) 13.4 (0.4) Sodium (mmol/l) mean (SE) 137(0.7) Potassium (mmol/l) mean, (SE) 4.3 (0.1) Urea (mmol/l) mean (SE) 10.1 (0.8) Creatinine (umol/l) mean, (SE) 127(6.0)
79
Figure 11: Flow diagram of patients in the study.
CRT=Cardiac resynchronisation therapy, LV=Left ventricular, RVOT=Right ventricular outflow tract,
24 patients recruited 1 declined CRT 1 diagnosed with renal ca 22 patients for implantation
1 failed LV lead 19 LV leads
10 CRT-D, 11 CRT- P, Upgrades of RV pacing 21 patients with successful CRT/Multisite pacing 21 patients available for 6 month follow up 20 patients available for 12month follow up 1 death 2 RVOT leads
80