Cost-Effectiveness Analysis of Quadripolar Versus Bipolar Left Ventricular Leads for Cardiac Resynchronization Defibrillator Therapy in a Large, Multicenter UK Registry

Objectives The objective of this study was to evaluate the cost-effectiveness of quadripolar versus bipolar cardiac resynchronization defibrillator therapy systems. Background Quadripolar left ventricular (LV) leads for cardiac resynchronization therapy reduce phrenic nerve stimulation (PNS) and are associated with reduced mortality compared with bipolar leads. Methods A total of 606 patients received implants at 3 UK centers (319 Q, 287 B), between 2009 and 2014; mean follow-up was 879 days. Rehospitalization episodes were costed at National Health Service national tariff rates, and EQ-5D utility values were applied to heart failure admissions, acute coronary syndrome events, and mortality data, which were used to estimate quality-adjusted life-year differences over 5 years. Results Groups were matched with regard to age and sex. Patients with quadripolar implants had a lower rate of hospitalization than those with bipolar implants (42.6% vs. 55.4%; p = 0.002). This was primarily driven by fewer hospital readmissions for heart failure (51 [16%] vs. 75 [26.1%], respectively, for quadripolar vs. bipolar implants; p = 0.003) and generator replacements (9 [2.8%] vs. 19 [6.6%], respectively; p = 0.03). Hospitalization for suspected acute coronary syndrome, arrhythmia, device explantation, and lead revisions were similar. This lower health-care utilization cost translated into a cumulative 5-year cost saving for patients with quadripolar systems where the acquisition cost was <£932 (US $1,398) compared with bipolar systems. Probabilistic sensitivity analysis results mirrored the deterministic calculations. For the average additional price of £1,200 (US $1,800) over a bipolar system, the incremental cost-effective ratio was £3,692 per quality-adjusted life-year gained (US $5,538), far below the usual willingness-to-pay threshold of £20,000 (US $30,000). Conclusions In a UK health-care 5-year time horizon, the additional purchase price of quadripolar cardiac resynchronization defibrillator therapy systems is largely offset by lower subsequent event costs up to 5 years after implantation, which makes this technology highly cost-effective compared with bipolar systems.

C ardiac resynchronization therapy (CRT) is an efficacious and costeffective (1) treatment for patients with symptomatic heart failure with poor left ventricular (LV) function and prolonged QRS duration (2)(3)(4). Despite improvements in implantation delivery equipment and accumulation of user experience over the past 2 decades, approximately 30% of patients do not derive symptomatic benefit (5,6). Post-implantation complications such as high capture thresholds, phrenic nerve stimulation (PNS), lead displacement, and infection reduce the effectiveness of this therapy (7)(8)(9)(10). The recent introduction of multipolar (quadripolar) LV leads has demonstrated a reduction in PNS through more proximal pole reprogramming, the presence of sustained lower capture thresholds, and easy deliverability (11).
However, new technology is usually provided at a higher purchase price than the conventional standard of care, which means that cost-effectiveness and affordability must be considered (12). Furthermore, the need for £22 billion in savings by 2020 in the United Kingdom (13) and an increased focus on efficiency as a result (14) further highlight the impor- We set out to assess the cost-effectiveness of quadripolar LV leads compared with bipolar LV leads in patients implanted with a cardiac resynchronization defibrillator therapy device (CRTD) within our previously published registry. We analyzed longer-term health-care utilization costs in terms of hospitalizations that occurred within the 5-year follow-up period to investigate whether the higher purchase price of this new technology was offset by expected reductions in cost arising from a reduction in hospitalizations. We also used mortality, acute coronary syndrome, and heart failure hospitalization data to estimate quality-adjusted life-year (QALY) differences. complications, and all-cause mortality (17).

METHODS
For the purposes of the current study, hospitalization episodes for each patient in the clinical registry were reviewed and assigned to the following categories based on diagnosis: acute coronary syndrome (ACS), arrhythmia, heart failure hospitalizations, infection requiring system explantation and reimplantation, generator replacement, and revision of any lead. These were compared between patients implanted with a CRTD incorporating a quadripolar LV lead versus those with a bipolar LV lead.  Cost-Effectiveness of Quadripolar vs. Bipolar CRTD F E B R U A R Y 2 0 1 7 : 1 0 7 -1 6 hospitalization ( Table 1). There was no extrapolation of data or event rates beyond the 5-year follow-up after implantation. Event rates were those that were observed to have occurred in each year; we did not derive transition probabilities that could be used for a Markov model. All events were counted, and some events occurred more than once in individual patients. A probabilistic sensitivity analysis was also undertaken to help understand the impact of parameter uncertainty and determine the probability that quadripolar CRTD was cost-effective. Probabilistic analysis was conducted by inputting data as probability (beta) distributions rather than point estimates and randomly sampling 1,000 values from these distributions. This was performed for all hospitalization episodes in addition to mortality data from our previous work (17). Comparative purchase costs were estimated between the quadripolar Quar-  Table 1. COSTS. National tariff "enhanced tariff option" prices for 2015 to 2016 (18) were applied to ACS hospitalization, arrhythmia hospitalization, heart failure hospitalization, and lead revision procedures. The base tariff price was multiplied by the local cost factor (market forces factor) for each NHS hospital that implants CRT devices, and the mean of these values was used in the model. Table 1 shows the mean unit cost data used in the calculations per hospitalization, including local cost factors. Online Table 2 shows the equivalent costs in US dollars using a simple conver-   Only the mortality difference used in our previous report (17), utility loss attributable to ACS events, and utility loss attributable to heart failure hospitalizations were used to assess QALY differences between    Table 3. Moreover, the proportion of patients hospitalized at least once was also significantly lower in those implanted with a quadripolar compared with a bipolar lead (42.6% vs. 55.4%, respectively; p ¼ 0.002), as shown in Table 4.     Cost-Effectiveness of Quadripolar vs. Bipolar CRTD lead revisions were similar between the groups (p ¼ NS). Each hospitalization, irrespective of cause, was counted as a separate event ( Table 3); these values were multiplied by the appropriate tariff ( Table 1) to produce health-care utilization costs for each group over the 5-year period. Table 4 represents the proportion of patients implanted with either quadripolar or bipolar leads who had at least 1 admission for the listed reasons.
The absolute values for the hospitalization causes are therefore less than in Table 3 (Figure 3). The cost saving was up to £1,000 ($1,500) for purchasing a quadripolar system for the same price as a bipolar system ( Table 5). Beyond £932 ($1,398), the additional ICER was up to £20,288 ($30,432) ( Figure 3). Figure 2 shows the impact of varying a range of input parameters by AE 95% confidence intervals. The analysis was most sensitive to the utility of patients with heart failure, because death resulted in a loss of 0.8808 QALYs in each patient who died. All resulting ICERs remained <£4,000 per QALY gained.
In the probabilistic sensitivity analysis, quadripolar CRTD was 97.1% likely to be cost-effective at £20,000 per QALY gained and 99.3% likely to be costeffective at £30,000 per QALY gained (Figure 4).

DISCUSSION
This is the first comprehensive health economic analysis to use real-world UK clinical data from hospitalization events and mortality to produce an accurate comparison of cumulative cost differences between implanting quadripolar versus bipolar CRTD systems.
The main findings were as follows:   Values are n (%).
CI ¼ confidence interval; other abbreviations as in Table 3.
Behar et al.  (11,26). Rates of intraprocedural lead complications appear lower than with conventional bipolar leads (27). Reduction or even elimination in PNS during medium-term follow-up provides invaluable utility in CRT delivery (9,15). We have recently shown a reduction in all-cause mortality associated with quadripolar leads compared with a bipolar lead (17). Furthermore, rates of reintervention for lead repositioning were lower in those implanted with a quadripolar compared with a bipolar lead (2% vs. 5.2%; p ¼ 0.03), and the radiation dose during implantation was almost one-half (1,028 cGy$cm 2 vs. 1,950 cGy$cm 2 ; p < 0.001).
The lower rates of hospitalization associated with a quadripolar lead in the current study could be driven by the improved efficacy in CRT delivery (attributable to PNS reduction and fewer reinterventions for lead displacement). Our previous study (17) also demonstrated lower implantation capture energy with quadripolar than with bipolar leads (0.95 mJ vs. 1.08 mJ; p ¼ 0.003). Pacing systems consistently delivering higher-output voltages to capture the LV will have a reduced longevity (28), and this could explain the current findings of a    per QALY gained. Abbreviations as in Figure 2.  applied NHS tariffs to them to determine the actual charge and cost-effectiveness. This was an in-study cost-effectiveness analysis, not an extrapolation to a lifetime horizon. We therefore did not assume event rates and did not model beyond the time for which we had gathered follow-up data. We did not perform a Markov model. Wider societal benefit was also not taken into account, which might be a further limitation.
As might be expected, the incremental acquisition cost of quadripolar technology is a strong determinant of the overall incremental costeffectiveness of the 2 therapies. We therefore made an estimate of base cost and performed an analysis either side of the additional purchase cost to account for the variation in procurement acquisition costs. With respect to QALYs, the mortality difference was the strongest driver of the QALY gain associated with quadripolar CRTD. There was a significant difference in the proportions of patients with ischemic heart disease and those not in sinus rhythm (with more such patients in the bipolar group); however, this was corrected for in the multivariate analysis, and mortality remained significantly different.