Search

BACKGROUND Cardiac resynchronization therapy (CRT) using bi-ventricular pacing (BVP) is effective in patients with heart failure, left bundle branch block (LBBB), and reduced left ventricular function. Left bundle branch area pacing (LBBAP) has been reported as an alternative option for CRT. OBJECTIVE The purpose of this study was to assess the feasibility and outcomes of LBBAP in patients who failed conventional BVP because of coronary venous (CV) lead complications or who were nonresponders to BVP. METHODS At 16 international centers, LBBAP was attempted in patients with conventional CRT indication who failed BVP because of CV lead complications or lack of therapeutic response to BVP. Heart failure hospitalization (HFH) and death, echocardiographic outcomes, procedural data, pacing parameters, and lead complications including CV lead failure are reported. RESULTS LBBAP was successfully performed in 200 patients (CV lead failures 156; nonresponders 44) (age 68 +/- 11 years; female 35%; LBBB 55%; right ventricular pacing 23%; ischemic cardiomyopathy 28%; nonischemic cardiomyopathy 63%; left ventricular ejection fraction [LVEF]

The introduction of dedicated tools for pacing and defibrillator lead extraction has resulted in relatively high success and low complication rates. The confidence this elicits has broadened the indications from device infections to non-functional or redundant leads and the latter make up an increasing share of extraction procedures. Proponents of extracting these leads point to the higher complication burden of lead extraction in patients with longstanding abandoned leads when compared one-to-one with extraction when these leads become redundant. However, this does not translate into better patient outcomes on a population level: complications are rare with properly abandoned leads and thus most patients will never be subjected to an extraction procedure and the ensuing complications. Therefore, not extracting redundant leads minimises the risk for the patients and avoids many expensive procedures.

Background Venous access for cardiac implantable electronic devices (CIED) is commonly performed by cephalic venous cut down, or axillary or subclavian vein puncture. The latter technique carries a risk of complications such as pneumothorax or lead crush. Cephalic venous cut down is free of these complications but often less successful due to technical difficulties. We report a highly successful, simplified cephalic venous access with a modified Seldinger technique. Methods We prospectively studied 221 patients undergoing de novo implantation of a one, two, or three lead device system performed over a 1-year period at our center, and assessed the efficacy of the cephalic vein access including the number of leads successfully placed for each procedure. Results In 83% of patients undergoing CIED implantation, a suitable cephalic vein was present. In respectively 98% and 99% of patients undergoing single- or dual-chamber CIED implantation, lead placement could be performed exclusively via the cephalic vein and in 72% of cardiac resynchronization therapy implants, all three leads were placed via cephalic access. Conclusion A novel, technically simple cephalic venous catheterization technique provides high success rates for any CIED lead implantation.

Right ventricular pacing (RVP) induces abnormal electrical activation and asynchronous ventricular contraction and leads to pacing induced cardiomyopathy (PICM) in 10%-20% of patients. Cardiac resynchronization therapy (CRT) utilizing biventricular pacing (BVP) is the recommended treatment. Left bundle branch pacing (LBBP) is a novel physiological pacing technique which may serve as an alternative to CRT. This study assessed feasibility and outcomes of LBBP delivered CRT in patients with PICM.Total 20 consecutive patients with PICM who received an upgrade of their pacemaker to LBBP were prospectively studied. Acute success rate, complications, functional and echocardiographic response, and hospitalization for heart failure within 6 months from implantation were evaluated.LBBP was successfully delivered in all patients. Median duration of RVP before upgrade to LBBP was 3.8 years and the RVP was 99%. LBBP resulted in significant QRS narrowing (from 193 ± 18 ms to 130 ± 17 ms [p .001]), improvement in LVEF (from 32% ± 6 % to 47% ± 8% [p .001]) and NYHA class (from 2.8 ± 0.4 to 1.4 ± 0.5 [p .001]) at 6 months. No LBBP-related complications occurred. No patients were hospitalized for heart failure or died.LBBP is feasible and safe in delivering CRT in PICM. Preliminary analyses demonstrated significant electrical resynchronization and favorable improvement in LV function and NYHA functional class at short term follow-up. Data needs to be validated in large randomized controlled trials.

BACKGROUND Cardiac resynchronization therapy (CRT) based on the conventional biventricular pacing (BiV-CRT) technique sometimes results in broad QRS complex and suboptimal response. OBJECTIVE We aimed to assess the feasibility and outcomes of CRT based on left bundle branch area pacing (LBBAP, in lieu of the right ventricular lead) combined with coronary venous left ventricular pacing in an international multicenter study. METHODS LBBAP-optimized CRT (LOT-CRT) was attempted in nonconsecutive patients with CRT indications. Addition of the LBBA (or coronary venous) lead was at the discretion of the implanting physician, who was guided by suboptimal paced QRS complex, and/or on clinical grounds. RESULTS LOT-CRT was successful in 91 of 112 patients (81%). The baseline characteristics were as follows: mean age 70 +/- 11 years, female 22 (20%), left ventricular ejection fraction 28.7% +/- 9.8%, left ventricular end-diastolic diameter 62 +/- 9 mm, N-terminal pro-B-type natriuretic peptide level 5821 +/- 8193 pg/mL, left bundle branch block 47 (42%), nonspecific intraventricular conduction delay 25 (22%), right ventricular pacing 26 (23%), and right bundle branch block 14 (12%). The procedure characteristics were as follows: mean fluoroscopy time 27.3 +/- 22 minutes, LBBAP capture threshold 0.8 +/- 0.5 V @ 0.5 ms, and R-wave amplitude 10 mV. LOT-CRT resulted in significantly greater narrowing of QRS complex from 182 +/- 25 ms at baseline to 144 +/- 22 ms (P < .0001) than did BiV-CRT (170 +/- 30 ms; P < .0001) and LBBAP (162 +/- 23 ms; P < .0001). At follow-up of >= 3 months, the ejection fraction improved to 37% +/- 12%, left ventricular end-diastolic diameter decreased to 59 +/- 9 mm, N-terminal pro-B-type natriuretic peptide level decreased to 2514 +/- 3537 pg/mL, pacing parameters were stable, and clinical improvement was noted in 76% of patients (New York Heart Association class 2.9 vs 1.9). CONCLUSION LOT-CRT is feasible and safe and provides greater electrical resynchronization as compared with BiV-CRT and could be an alternative, especially when only suboptimal electrical re synchronization is obtained with BiV-CRT. Randomized controlled trials comparing LOT-CRT and BiV-CRT are needed.

Objectives The purpose of this study was to investigate whether stimulation at the left ventricular (LV) septum (LVs), alone or in combination with another site, could be an alternative way to apply cardiac resynchronization therapy (CRT) that avoids the coronary sinus and phrenic nerve stimulation and may create more physiological sequence of activation. Background In CRT, biventricular pacing is commonly performed from the right ventricle (RV) and the epicardium of the LV lateral wall (LVlat). In the left bundle branch block (LBBB), half of the electrical delay occurs due to impulse conduction across the septum. Methods Experiments were performed in 13 dogs with LBBB, 7 of them with chronic myocardial infarction (LBBB + MI). Pacing leads were positioned in the right atrium, RV, LVs, and at the LVlat epicardium. LV pump function was measured using conductance catheter and synchrony of electrical activation of the ventricles using epicardial mapping and from surface electrocardiogram. In 12 CRT patients, LV pump function was measured during temporary RV + LVs pacing and compared to RV + LVlat and RV + LVlat endo pacing. Results In the animals, electrical and hemodynamic benefits of LVs and RV + LVs pacing were comparable to those during conventional biventricular pacing and were comparable in LBBB and LBBB + MI hearts. Dispersion of repolarization was reduced by LVs stimulation, but not by LVlat pacing. In patients, hemodynamic benefits of RV + LVs, RV + LVlat and RV + LVlat endo pacing were similar. Conclusions The use of the LVs as LV pacing site in CRT improves synchronization and acute hemodynamics comparably to conventional biventricular pacing in dyssynchronous canines and in patients. In addition, LVs stimulation may reduce dispersion of repolarization compared to epicardial pacing.

Endocardial left ventricular (LV) pacing for cardiac resynchronization therapy (CRT) has been proposed as an alternative to traditional LV transvenous epicardial pacing with equal or superior cardiac performance. The risks of cerebral thromboembolism and possible interference with mitral valve function moderate its clinical application.The purpose of this study was to investigate cerebral thromboembolic complications after LV endocardial lead placement. Mitral regurgitation (MR) was the secondary outcome measure.CRT candidates with a failed coronary sinus approach or nonresponders to conventional CRT underwent endocardial LV lead implantation (45 atrial transseptal, 6 transapical). Coumarin was prescribed with a targeted international normalized ratio between 3.5 and 4.5. Patient records were checked and general practitioners were contacted regarding cerebral thromboembolic complications. MR was evaluated by echocardiography at baseline and after 6 months.In 7 patients, 6 ischemic strokes and 2 transient ischemic attacks occurred, corresponding to 6.1 thromboembolic events per 100 patient-years (95% confidence interval 3.4-15.8). One patient refused hospital admission; all other patients had a subtherapeutic anticoagulation level at the time of the event. No major bleeding complications occurred. There was no change in the grade of MR (grade 2, P = .727) after 6 months.Endocardial LV lead placement in patients with advanced heart failure is associated with thromboembolic risk. However, all but 1 patient had a subtherapeutic level of anticoagulation. Endocardial LV lead placement is not associated with aggravation of MR.

Background— In cardiac resynchronization therapy (CRT), optimization of left ventricular (LV) stimulation timing is often time consuming. We hypothesized that the QRS vector in the vectorcardiogram (VCG) reflects electric interventricular dyssynchrony, and that the QRS vector amplitude (VA QRS ), halfway between that during left bundle branch block (LBBB) and LV pacing, reflects optimal resynchronization, and can be used for easy optimization of CRT. Methods and Results— In 24 canine hearts with LBBB (12 acute, 6 with heart failure, and 6 with myocardial infarction), the LV was paced over a wide range of atrioventricular (AV) delays. Surface ECGs were recorded from the limb leads, and VA QRS was calculated in the frontal plane. Mechanical interventricular dyssynchrony (MIVD) was determined as the time delay between upslopes of LV and right ventricular pressure curves, and systolic function was assessed as LV dP/dt max . VA QRS and MIVD were highly correlated ( r =0.94). The VA QRS halfway between that during LV pacing with short AV delay and intrinsic LBBB activation accurately predicted the optimal AV delay for LV pacing (1 ms; 95% CI, –5 to 8ms). Increase in LV dP/dt max at the VCG predicted AV delay was only slightly lower than the highest observed ∆LV dP/dt max (–2.7%; 95% CI, –3.6 to –1.8%). Inability to reach the halfway value of VA QRS during simultaneous biventricular pacing (53% of cases) was associated with suboptimal hemodynamic response, which could be corrected by sequential pacing. Conclusions— The VA QRS reflects electric interventricular dyssynchrony and accurately predicts optimal timing of LV stimulation in canine LBBB hearts. Therefore, VCG may be useful as a reliable and easy tool for individual optimization of CRT.

Background— Several studies suggest that patients with ischemic cardiomyopathy benefit less from cardiac resynchronization therapy. In a novel animal model of dyssynchronous ischemic cardiomyopathy, we investigated the extent to which the presence of infarction influences the short-term efficacy of cardiac resynchronization therapy. Methods and Results— Experiments were performed in canine hearts with left bundle branch block (LBBB, n=19) and chronic myocardial infarction, created by embolization of the left anterior descending or left circumflex arteries followed by LBBB (LBBB+left anterior descending infarction [LADi; n=11] and LBBB+left circumflex infarction [LCXi; n=7], respectively). Pacing leads were positioned in the right atrium and right ventricle and at 8 sites on the left ventricular (LV) free wall. LV pump function was measured using the conductance catheter technique, and synchrony of electrical activation was measured using epicardial mapping and ECG. Average and maximal improvement in electric resynchronization and LV pump function by right ventricular+LV pacing was similar in the 3 groups; however, the site of optimal electrical and mechanical benefit was LV apical in LBBB hearts, LV midlateral in LBBB+LCXi hearts and LV basal-lateral in LBBB+LADi hearts. The best site of pacing was not the site of latest electrical activation but that providing the largest shortening of the QRS complex. During single-site LV pacing the range of atrioventricular delays yielding ≥70% of maximal hemodynamic effect was approximately 50% smaller in infarcted than noninfarcted LBBB hearts ( P Conclusions— Cardiac resynchronization therapy can improve resynchronization and LV pump function to a similar degree in infarcted and noninfarcted hearts. Optimal lead positioning and timing of LV stimulation, however, require more attention in the infarcted hearts.

Background: Hip fractures are associated with high morbidity. Pressure ulcer formation after hip surgery is often related to delayed patient mobilization. The objectives of this study were to determine whether time-to-surgery affects development of pressure ulcers postoperatively and, thus, length of hospital stay. Patients and Methods: We performed a retrospective analysis of consecutive hip fracture patients, aged 60 years and above, who underwent surgery between 1995 and 2001. The primary outcome was in-hospital development of pressure ulcers. The secondary outcome measure was the overall length of hospital stay. Analyses were adjusted for relevant confounders. Results: Of the 722 patients enrolled, 488 patients (68%) received surgery at 12 h after admission. Approximately 30% (n = 214) developed pressure ulcers during admission, whilst 19% of patients operated within 12 h of admission developed pressure ulcers. Time-to-surgery was an independent predictor of both development of pressure ulcers (OR = 1.7, 95% confidence interval [CI] = 1.2–2.6; p = 0.008) and length of hospital stay (11.3 vs 13.3 days in the early and the late surgery group, respectively, p = 0.050). Furthermore, development of pressure ulcers was associated with prolonged postoperative hospital stay (19.5 vs 11.1 days for patients with and without pressure ulcers, respectively, p = 0.001) Interpretation: In hip fracture patients, time-to-surgery was an independent predictor of both postoperative pressure ulcer development and prolonged hospital stay. These data suggest that the implementation of an early surgery protocol following admission for hip fractures may reduce both the postoperative complications and overall hospital stay.

OBJECTIVES The aim of this study was to investigate the mid-term haemodynamic and clinical results after aortic valve replacement (AVR) using the Sorin Freedom Solo (SFS) stentless bioprosthesis, compared with the standard Carpentier Edwards Perimount (CEP) stented bioprosthesis. METHODS In this retrospective cohort study of prospectively collected data, 116 patients were included in the SFS group (53 males; median age 74 years, range 56-85 years), and 122 patients in the CEP group (85 males; median age 73 years, range 43-88 years) between July 2007 and January 2013. Echocardiography was performed at 6 weeks after surgery in our centre, and the most recent echocardiography (in our centre or in referring cardiology departments) was requested. Between September 2013 and April 2014, all patients were called by the same researcher to gain clinical follow-up data. RESULTS Mid-term mortality was 16.4% in the SFS group (19 patients) and 21.3% in the CEP group (26 patients); (P = 0.3). The mean transvalvular gradient was 7.4 ± 3.1 mmHg in the SFS group, and 11.6 ± 3.2 mmHg in the CEP group at 6 weeks postoperatively (P < 0.001). When stratified by labelled valve size, mean gradients were significantly lower in the SFS group for every size (P ≤ 0.03). After 3.3 ± 1.4 years of follow-up, the mean gradient was still significantly lower in the SFS group than that in the CEP group (P < 0.001). Clinical follow-up showed relatively low complication rates. CONCLUSION These data suggest that the Sorin Freedom Solo stentless bioprosthesis is as safe as the Carpentier Edwards bioprosthesis, and provides better short- and mid-term haemodynamic performance than the Carpentier Edwards bioprosthesis.

A 76-year-old man with a 1-week history of continuous chest pain presented to the ED with rapidly progressive dyspnea on exertion. Transthoracic echocardiography revealed a large, sharply demarcated interventricular septal defect with a turbulent left-to-right transseptal flow, shown in videos.

Background— Studies in canine hearts with acute left bundle branch block (LBBB) showed that endocardial left ventricular (LV) pacing improves the efficacy of cardiac resynchronization therapy (CRT) compared with conventional epicardial LV pacing. The present study explores the efficacy of endocardial CRT in more compromised hearts and the mechanisms of such beneficial effects. Methods and Results— Measurements were performed in 22 dogs, 9 with acute LBBB, 7 with chronic LBBB combined with infarction (embolization; LBBB plus myocardial infarction, and concentric remodeling), and 6 with chronic LBBB and heart failure (rapid pacing, LBBB+HF, and eccentric remodeling). A head-to-head comparison was performed of the effects of endocardial and epicardial LV pacing at 8 sites. LV activation times were measured using ≈100 endocardial and epicardial electrodes and noncontact mapping. Pump function was assessed from right ventricular and LV pressures. Endocardial CRT resulted in better electric resynchronization than epicardial CRT in all models, although the benefit was larger in concentrically remodeled LBBB plus myocardial infarction than in eccentrically remodeled LBBB+HF hearts (19% versus 10%). In LBBB and LBBB+HF animals, endocardial conduction was ≈50% faster than epicardial conduction; in all models, transmural impulse conduction was ≈25% faster when pacing from the endocardium than from the epicardium. Hemodynamic effects were congruent with electric effects. Conclusions— Endocardial CRT improves electric synchrony of activation and LV pump function compared with conventional epicardial CRT in compromised canine LBBB hearts. This benefit can be explained by a shorter path length along the endocardium and by faster circumferential and transmural impulse conduction during endocardial LV pacing.

Background— We investigated the benefits of the more physiological activation achieved by left ventricular (LV) endocardial pacing (ENDO) as compared with conventional epicardial (EPI) LV pacing in cardiac resynchronization therapy. Methods and Results— In 8 anesthetized dogs with experimental left bundle-branch block, pacing leads were positioned in the right atrium, right ventricle, and at 8 paired (EPI and ENDO) LV sites. Systolic LV pump function was assessed as LVdP/dtmax and stroke work and diastolic function as LVdP/dtmin. Electrical activation and dispersion of repolarization were determined from 122 epicardial and endocardial electrodes and from analysis of the surface ECG. Overall, ENDO-biventricular (BiV) pacing more than doubled the degree of electrical resynchronization and increased the benefit on LVdP/dtmax and stroke work by 90% and 50%, respectively, as compared with EPI-BiV pacing. During single-site LV pacing, the range of AV intervals with a >10% increase in LV resynchronization (79�31 versus 32�24 ms, P Conclusions— Data from this acute animal study indicate that the use of an endocardial LV pacing electrode may increase the efficacy of resynchronization therapy as compared with conventional epicardial resynchronization therapy.

Amiodarone and sotalol are frequently used in the treatment of atrial fibrillation. However, oral and intravenous (IV) therapy with these drugs has suboptimal efficacy and is associated with serious extracardiac side effects. We hypothesized that intrapericardial (IPC) delivery produces antiarrhythmic effects at lower plasma drug concentrations than IV delivery. Goats (n = 27) were randomised into 5 groups receiving either IPC vehicle, amiodarone (IV or IPC) or dl-sotalol (IV or IPC). Epicardial and endocardial atrial effective refractory period and atrial response to burst pacing (rapid atrial response, RAR) were assessed before and after 3 hours of drug infusion at 2 mg.kg.h. IPC delivery produced steeply decreasing drug concentrations from epicardium to endocardium in both atria and ventricles. Plasma drug concentrations were significantly lower in IPC than in IV groups. IPC amiodarone and sotalol reduced epicardial RAR inducibility (-74% +/- 20% and -66% +/- 30%, respectively) compared with IV delivery (-11% +/- 17% and -17% +/- 28%, respectively; P < 0.05). Endocardial RAR inducibility was only reduced in the IPC amiodarone group (-70% +/- 17%, P < 0.05). In conclusion, IPC delivery of amiodarone and sotalol increases atrial drug concentration and antiarrhythmic effects at reduced plasma drug concentrations. These potential benefits are particularly prominent for IPC delivered amiodarone.

Amiodarone (AM) is one of the most effective anti-arrhythmic drugs to prevent postoperative atrial fibrillation (AF). However, due to systemic side effects, prophylactic AM therapy is only appropriate for patients at high risk for postoperative AF. This study addressed the hypothesis that local epicardial delivery of AM produces therapeutic myocardial drug concentrations, while systemic levels remain low. Goats (n=14) were instrumented with right atrial epicardial patch electrodes, and a PEG-based hydrogel with AM (1 mg/kg bw) (n=9) or without drug (n=5) was applied to the right atrial epicardium. AF inducibility was assessed up to 28 days in awake goats by applying burst stimuli. Myocardial and plasma AM concentrations were analysed by HPLC. AM-hydrogel produced higher subepicardial than subendocardial drug concentrations, which both remained therapeutic up to 21 days after hydrogel application (fig. 1a ). In this period, AF inducibility was significantly lower in the AM-hydrogel group compared to hydrogel alone (fig. 1b ; p Figure 1a. Amiodarone Distribution in Right Atrium (n=5) Figure 1b. AF Inducibility After Hydrogel Application (mean + SD)

This article illustrates the highly interesting possibilities of cardiac resynchronization therapy (CRT) yet also discusses several problems for optimal implementation. The authors have presented possible strategies to improve CRT by designing and using better measurements of mechanical dyssynchrony or discoordination, leading to better patient selection. In addition, application of CRT can be improved by developing techniques to reach the optimal pacing site(s) easier and algorithms for easier programming of optimal atrioventricular delay and interventricular interval. Also, application of CRT in patients with less severe heart failure deserves further attention. Finally, the mechanism of ventricular interaction may open new fields of application of left ventricle–based pacing in patients who have heart failure, even with a narrow QRS complex.

Background: Amiodarone causes less ventricular transmural dispersion of repolarization (TDR) and proarrhythmia than sotalol. Effects of both drugs on atrial TDR, however, have not yet been well defined. We applied intrapericardial (IPC) drug delivery to produce in vivo transmural concentration gradients. We assessed the hypotheses that IPC amiodarone causes less atrial TDR than IPC sotalol and that IPC amiodarone is more effective than IPC sotalol in lowering atrial fibrillation (AF) inducibility. Methods: Goats (n = 25) were randomised into 5 groups receiving either IPC dextrose, amiodarone (IV or IPC) or d,l-sotalol (IV or IPC). Drugs were infused for 3 hours at a constant rate of 2 mg/kg per hour. Quadripolar catheters were placed on the epi- and endocardial surfaces of the right atrium. Atrial effective refractory period (AERP) and AF inducibility were measured by single premature stimuli at 4 times the diastolic threshold. Atria were freeze-microtomed parallel to the epicardial surface and analysed by HPLC. Results: IPC drug delivery produced high epicardial concentrations (fig.1a ) with significantly decreased epicardial AF inducibility. A similar decrease in endocardial AF inducibility was only observed in the IPC amiodarone group (fig.1b ). TDR (defined as endo-minus epicardial AERP) became negative after IPC sotalol (from 13.0 ± 3.2 to -30.7 ± 7.3 ms, p < 0.05). In the other groups TDR did not change significantly and remained positive. Conclusions: IPC delivery boosts epicardial drug concentration and efficacy. For amiodarone, this also increases endocardial efficacy. IPC sotalol, however, inverts TDR, which may explain its inferior endocardial efficacy. Figure 1a. Concentrations in right atrium (mean + SD) Figure 1b. Chang in AF inducibility (mean + SD)

ObjectiveAutologous platelet clots serve as slow-release delivery systems for platelet-derived growth factors and cytokines. Their application to the pericardial sac might facilitate salvage and repair of ischemically injured myocardium. However, little is known about platelet clot stability in the pericardial sac. We investigated the stability of platelet clots in vitro and after administration to the pericardial sac in pigs and patients.MethodsIn 5 Yorkshire-Landrace pigs and 10 patients, in vitro manufactured autologous platelet gel (Medtronic Magellan Platelet Separator) and platelet-rich fibrin (Vivolution Vivostat System) were administered to the pericardial sac for 30 minutes. Two antifibrinolytics (tranexamic acid and aprotinin) were tested for their capacity to stabilize autologous platelet gel. In vitro clots, incubated at 37°C for 48 hours, served as controls. Clot weight was measured before and after administration.ResultsIn vitro, autologous platelet gel clots of either formula liquefied almost entirely within 60 minutes whereas platelet-rich fibrin clots remained intact. In the pig, platelet clot weight decreased to 16.7% ± 7.8% (P < .05) and 66.4% ± 3.2% (P < .05) of initial clot weight for autologous platelet gel and platelet-rich fibrin, respectively. Addition of antifibrinolytics to autologous platelet gel did not reduce clot degradation significantly. In patients, autologous platelet gel and platelet-rich fibrin clot weight remained 9.0% ± 1.5% (P < .05) and 73.7% ± 2.6% (P < .05) of initial clot weight, respectively.ConclusionsAutologous platelet gel is unstable both in vitro and in vivo, whereas platelet-rich fibrin remains intact in vitro and, compared with autologous platelet gel, is less subject to degradation in pigs and in patients.