OUP user menu

Management of recalled implantable cardioverter-defibrillator leads at generator replacement: a decision analysis model for Fidelis leads

Haran Burri, Christophe Combescure
DOI: http://dx.doi.org/10.1093/europace/eut425 1210-1217 First published online: 27 January 2014


Aims Fidelis defibrillator leads have been recalled in 2007 because of a high fracture rate, exposing patients to inappropriate shocks, and to re-interventions. There are currently no guidelines on the management of patients with functional Fidelis leads. The dilemma is of particular relevance when the patient undergoes a generator replacement, as this offers an opportunity to perform concomitant lead revision. Our aim was to evaluate long-term outcome with different strategies of Fidelis lead management at generator replacement.

Methods and results A decision analysis model comparing generator replacement only with different additional lead revision strategies was constructed with yearly cycles for up to 10 years. Published data were used to calculate outcomes of mortality, inappropriate shocks due to lead fracture, and major complications related to lead revision. A quality-of-life analysis was also performed. There were no differences in total mortality between the different strategies. In case of a generator replacement only, the probability of Fidelis lead fracture requiring re-intervention cumulated to 36 and 49% at 5 and 10 years, respectively. Likewise, the risk of inappropriate shock due to lead fracture was 11 and 15%, respectively. A sensitivity analysis showed that all-cause mortality rates did not affect the results. Lead revision was also preferable in the quality-of-life analysis.

Conclusion In patients with a functional Fidelis lead at generator replacement, lead revision (with or without lead extraction) should be seriously considered to avoid re-intervention for subsequent lead fracture and to reduce inappropriate shocks.

  • Implantable cardioverter-defibrillator
  • Defibrillator lead
  • Complications
  • Lead dysfunction
  • Lead extraction
  • Decision analysis

What's new?

  • Our decision analysis model indicates that Fidelis lead revision (with or without extraction of the recalled lead) should be seriously considered at generator change.

  • Lead revision is also the preferred strategy according to the quality-of-life analysis.

  • Overall mortality has little impact on the decision analysis.


The Medtronic Sprint Fidelis lead is a thin-bodied implantable cardioverter-defibrillator (ICD) lead that had become popular because of its ease of handling, with more than a quarter of a million leads implanted worldwide.1 The lead was however withdrawn from the market in October 2007 because it was prone to fracture, resulting in inappropriate or inefficient shocks, or failure to pace. The rate of fracture approaches 17% at 5 years.2 In patients with functioning Fidelis leads, physicians are faced with the dilemma of whether they should perform prophylactic lead revision, especially when the patient is due for a generator replacement. There is currently no international consensus on how to deal with this situation. The decision is influenced by the probability of future lead fracture (which not only depends on the lead itself, but also on overall patient survival), procedural risk and operator experience (e.g. with lead extraction), as well as patient perception of different outcomes (e.g. inappropriate shock). A decision analysis model is well suited to analyse the different treatment options, especially as randomized-controlled data are unlikely to become available to guide decision-making in this context. There is now a wealth of published data regarding the Fidelis leads which may be used to construct a robust model. Decision analysis models dealing with recalled cardiac rhythm management devices have been reported, but do not specifically involve the Fidelis leads,3,4 focus on the costs,5 or only take into account survival as an outcome measure.4 Our aim was to evaluate different strategies with regard to Fidelis lead management at generator replacement, using published data and a decision analysis model that takes into account mortality, inappropriate shocks, major procedural complications, as well as quality of life.


Lead revision strategies

Four strategies for managing patients undergoing generator replacement were compared: (1) Generator replacement only, which avoids risks associated with lead revision but may expose the patient to inappropriate or inefficient therapy in case of future lead fracture (2) Adding a new pacing lead to replace the pace/sense (P/S) circuit, while continuing to use the high-voltage (HV) component of the Fidelis lead. Although Medtronic recommends against this in case of overt Fidelis lead fracture, there is no recommendation regarding doing this as a prophylactic measure. The rationale for this strategy is that failure occurs in the P/S component of the Fidelis lead in about 90–95% of cases.2,6,7 This strategy avoids inappropriate shocks associated with Fidelis lead fracture; however, revision will be necessary in case the HV component of the Fidelis lead fails. (3) Adding a new ICD lead without extracting the Fidelis lead. This increases the bulk of foreign body material and adds a small risk of lead–lead interactions. Also, in case of future issues (fracture of the new lead, infection, etc.), the complexity of the issue is increased due to more leads to deal with. The Medtronic 6947 Sprint Quattro model was chosen as it is frequently used in clinical practise (4) Extracting and replacing the Fidelis lead with a 6947 Sprint Quattro lead. This reduces the bulk of foreign body material in the patient's venous system as well as avoids potential lead–lead interactions, but exposes the patient to the risks of lead extraction.

Decision model structure

The decision analysis involves an initial phase at generator replacement modelled by a decision tree with the four above-mentioned strategies (Figure 1A). The initial strategy and the outcome of the generator replacement determines in which of the three states (defined by the lead model) the patient enters follow-up (Figure 1B). Follow-up was modelled over 10 years by a Markov model with yearly cycles. If lead fracture occurs, the patient is exposed to lead-related inappropriate shocks and will undergo in all cases a lead revision with a new Sprint Quattro ICD lead with or without extraction of the Fidelis lead. The full decision analysis model can be viewed in the Supplementary material online, the Appendix.

Figure 1

(A) Markov model structure for the initial phase at initial generator replacement with the different resulting states (Fidelis, patient has the initial Fidelis lead; Sprint Quattro, the patient has a Sprint Quattro ICD lead; PM lead + HV Fidelis, the patient has a pacemaker lead for the P/S circuit, while maintaining the HV component of the Fidelis lead. (B) Structure for analysis of yearly cycles, with the Fidelis state shown as an example here. The other states (Sprint Quattro and PM lead + HV Fidelis) are not shown here, but have identical structures with different rates of outcome (detailed in Table 1).

Parameters of the model

The data used for the main analysis as well as for sensitivity analyses are shown in Table 1. These data were extracted from publications relating to Fidelis leads and to device interventions, as well as from reports on lead performance on the Medtronic website. Survival was computed by using data from the ALTITUDE registry8 in the 28 869 ICD and cardiac resynchronization therapy-defibrillator patients at 4-year' follow-up (see the Supplementary material online, the Appendix for calculations). The annual risk of Fidelis lead fracture was calculated as 7.2%, based upon the failure rates between Years 4 and 5 of the largest series published to date relating to the 6949 model.2 The fracture rate has been reported to increase in the year following generator replacement (presumably due to damage to the lead in the pocket related to the procedure).10 Based on these data, we calculated the risk of Fidelis lead fracture to be 20.8% at Years 1 and 6, when generator replacement occurred (assuming battery longevity of 5 years). Data provided by the manufacturer were used to compute the risk of HV failure for Strategy 2 (see the Supplementary material online, the Appendix for the calculations). The risk of inappropriate shocks in the case of Fidelis lead fracture was taken as 30.6%, assuming that the Medtronic ‘Lead Integrity Alert’ algorithm is enabled in all the cases.2 This algorithm tracks changes in impedance and monitors short R–R intervals to allow early detection of lead fracture, and then triggers programming changes to reduce risk of inappropriate shock, as well as sets off an audible beep to alert the patient. Sensitivity analyses were performed with a higher risk of inappropriate shock in case the algorithm was absent2 (e.g. in the case of a generator from another manufacturer) or a lower risk in case the patient is placed on remote monitoring.12 For Strategy 2, no excess mortality was attributed to failure of the HV lead, as supported by recent data.16 The risk may be mitigated by strategies such as remote monitoring of HV lead impedance, assuming that this parameter portrays adequate HV lead function.

View this table:
Table 1

Parameters of the decision analysis and ranges for the sensitivity analyses

ParameterProbability (%)Range (%)Comments
Annual mortality (non-device-related)9.480.0–20.0Arbitrary values for sensitivity analysis (maximum of 20% as a generator replacement is probably futile in case of higher values)
Mortality the year following an infection and extraction21.09Data extracted from Figure 2 of the reference and adjusted for peri-operative mortality
Lead fracture
  First and sixth years20.810Higher rate reported following generator replacement
  Other years7.222.7–22.911Base rate calculated at Year 5 from references
 Sprint Quattro0.47
 HV Fidelis
  First and sixth years4.2Higher risk following generator replacement. HV data used for calculations of Strategy 2 only
  Other years1.50.5–3.0
 Pacing lead
  All years0A 100% lead survival rate was used to simplify calculations for Strategy 2
Inappropriate shock in case of lead fracture
 Fidelis30.6212.012–51.22Values differ according to the presence of lead integrity alert algorithm (only available on Medtronic devices) and use of remote monitoring
 Sprint Quattro21.77The P/S component is involved in 83% of Sprint Quattro fractures
Extraction of ICD lead in case of fracture50.00.0–100.0Arbitrary values, as local practice varies greatly
Life-threatening complication for lead revision without extraction0.513
Infection after generator replacement with or without addition of a lead1.714
Infection after intervention involving lead extraction2.813Higher rate than generator replacement only or lead revision due to more complex procedure
Life-threatening complication with lead extraction2.00.4–3.515Base rate calculated as the middle of the reported range
Mortality of lead extraction
 Lead dysfunction/prophylactic0.40.0–0.815Base rate taken as middle of reported range
 Extraction for infection2.29
  • HV, high voltage; P/S, pace–sense.

Although anecdotal reports exist of lead–lead interactions in patients with multiple right ventricular leads, the outcome has been reported to be favourable in ICD patients with an additional P/S or HV lead.1719 Therefore, no added risk was considered for the model in this context. No procedural deaths were considered for adding a pacemaker or ICD lead, in accordance with previous data,13,20 and no life-threatening major complications were attributed for a generator replacement only. Peri-procedural mortality and major complications of Fidelis lead extraction has been reported to be as low as 0% from high-volume centres.21 However, the duration of lead implantation was relatively short in this report (about 2 years on average), whereas Fidelis leads will currently have been implanted for at least 6 years and may thus be more difficult to extract. Data from several previously published series on lead extraction (that did not specifically involve Fidelis leads) reported life-threatening complications in 0.4–3.5% of patients, and death in 0–0.8%.15 For the purpose of the current analysis, the rate of life-threatening major complications (e.g. cardiac tamponade, vascular tear, pulmonary embolus, stroke, etc.22) was taken as 2.0% and procedural mortality rate for extracting non-infected leads was considered to be 0.4%. Peri-procedural mortality for patients with systemic infection was however considered to be higher at 2.2% as previously reported.9


Cohorts of 100 000 patients were simulated using R 2.15.1 software for Windows (R foundation for statistical computing, www.r-project.org). The main outcomes were mortality, inappropriate shocks related to lead failure, and re-intervention for lead revision, and life-threatening complications. To account for the uncertainty in the estimated value of parameters, probabilistic distributions based on published data were introduced in the model (see Table 1 and Supplementary material online, the Appendix). For each cohort, a set of parameters was randomly generated from these distributions and a total of 10 000 cohorts were simulated. The main outcomes were reported as cumulative percentages of events over follow-up. The 95% confidence intervals around the cumulative percentages were obtained over the 10 000 cohorts. Statistical comparisons between strategies were conducted using a Z-test based on the distributions of the outcomes obtained over the 10 000 cohorts. The significance level was set at 0.05. Moreover, to check the robustness of the results with regard to the key parameters of the model, a one-way sensitivity analysis was performed. The sensitivity analysis also facilitates clinical decision-making in individual patients whose profiles may differ from the parameters used in the main model. The decision analysis was completed by evaluating the effect of disutility (the opposite of utility) of adverse events (inappropriate shocks due to lead failure, re-intervention for lead dysfunction, and life-threatening complications of lead extraction). The disutility only affected the health state for the cycle during which the event occurred, as it has been shown that patients are usually able to cope well with adverse events; for example, quality of life is similar in patients with or without shocks within the last year (whereas it is reduced within 1–2 months after a shock).23


There was no difference in total mortality between the four strategies, as procedural mortality was negligible compared with overall all-cause mortality over follow-up (Table 2). The results for the other main outcome measures of inappropriate shocks, requirement for revision due to lead failure, and life-threatening complications related to lead extraction, are shown in Table 2 and Figure 2. Inappropriate shocks were significantly more frequent with Strategy 1 (by about 11 and 14% in absolute terms at 5 and 10 years, respectively) compared with the other strategies (P < 0.001). Although statistically significant (P < 0.001), the reduction in the inappropriate shocks with Strategy 2 was clinically irrelevant (absolute difference <1%) compared with the Strategies 3 and 4. Strategies 1 and 2 had a significantly greater risk of revision for lead failure than the other strategies at 5 and 10 years' follow-up (P ≤ 0.001 for all the comparisons). Life-threatening complications due to lead extraction were significantly greater with Strategy 4 compared with the other strategies (P ≤ 0.02), but absolute differences were small (<2%).

View this table:
Table 2

Outcome at 5 and 10 years' follow-up with the different strategies

Strategy 1
Generator replacement only
Strategy 2
PM lead + HV Fidelis
Strategy 3
New ICD lead
Strategy 4
New ICD lead with extraction
Main analysis (increased risk of lead fracture directly after generator replacement)
  5 years38.9 (37.4; 40.4)38.8 (37.3; 40.4)38.8 (37.3; 40.3)39.2 (37.7; 40.7)
  10 years62.5 (60.7; 64.4)62.5 (60.6; 64.3)62.4 (60.6; 64.3)62.7 (60.8; 64.5)
 Inappropriate shocksa
  5 years11.0 (7.4; 15.3)0.0 (0.0; 0.1)0.5 (0.1; 1.3)0.5 (0.1; 1.3)
  10 years14.9 (10.3; 20.1)0.1 (0.0; 0.2)0.7 (0.1; 2.1)0.7 (0.1; 2.1)
  5 years36.2 (27.9; 44.9)8.5 (5.6; 11.9)2.1 (0.6; 5.1)2.1 (0.6; 5.0)
  10 years49.0 (39.4; 58.4)13.3 (8.9; 18.3)3.3 (1.0; 8.2)3.3 (1.0; 8.1)
 Procedure-related infection
  5 years2.5 (1.6; 3.6)1.8 (1.1; 2.7)1.7 (1.0; 2.5)2.9 (1.2; 5.3)
  10 years2.8 (1.8; 4.0)1.9 (1.2; 2.9)1.7 (1.0; 2.6)2.9 (1.2; 5.3)
 Life-threatening complicationsb
  5 years0.5 (0.3; 0.8)0.6 (0.1; 1.2)0.5 (0.1; 1.6)2.0 (1.2; 3.0)
  10 years0.6 (0.4; 1.0)0.6 (0.2; 1.8)0.5 (0.1; 1.7)2.0 (1.2; 3.0)
Sensitivity analysis (constant risk of lead fracture over follow-up)
  5 years38.9 (37.4; 40.4)38.8 (37.4; 40.3)38.8 (37.3; 40.3)39.2 (37.7; 40.7)
  10 years62.5 (60.7; 64.3)62.5 (60.7; 64.3)62.5 (60.6; 64.3)62.7 (60.9; 64.5)
 Inappropriate shocksa
  5 years7.8 (5.1; 11.0)0.0 (0.0; 0.0)0.4 (0.1; 0.7)0.4 (0.1; 0.7)
  10 years11.0 (7.5; 15.2)0.0 (0.0; 0.1)0.6 (0.2; 1.2)0.6 (0.2; 1.2)
  5 years25.6 (18.7; 32.9)5.9 (3.7; 8.5)1.8 (1.7; 1.8)1.8 (1.7; 1.8)
  10 years36.2 (27.5; 44.9)9.3 (5.8; 13.2)2.9 (2.8; 2.9)2.9 (2.8; 2.9)
  • aEvents related to lead dysfunction.

  • bEvents related to lead revision with or without extraction.

  • Data are shown as percentages with 95% confidence intervals.

  • HV, high voltage; ICD, implantable cardioverter-defibrillator; P/S, pace–sense.

  • The results are for the main analysis assuming increased risk of the Fidelis lead fracture of 20.8% in the year of generator replacement, as well as for the sensitivity analysis of a constant annual fracture rate of 7.2%.

Figure 2

Cumulative risk of inappropriate shocks and re-intervention due to lead failure, and life-threatening complications related to lead extraction. P values refer to comparisons of the strategy with the highest risk compared with each of the other strategies.

The sensitivity analyses indicated that our model was robust. Data for variation in overall mortality for 5-year follow-up are shown in Figure 3 (results of analyses for the other variables, namely Fidelis fracture rate, rate of inappropriate shock, of life-threatening complications with lead extraction, and for 10-year follow-up are shown in the Supplementary material online, the Appendix). The rates of inappropriate shock due to lead fracture and requirement for lead revision with Strategy 1 stayed high despite the competing risk of mortality. As anticipated, risks of inappropriate shocks related to lead fracture and requirement for lead revision were affected by rates of Fidelis lead fracture for Strategy 1 (8–18 and 25–60%, respectively, over 5 years for the ranges reported in Table 1). Assuming a constant annual risk of Fidelis lead fracture over the entire follow-up (i.e. no increase during the year of generator replacement), the results still showed markedly greater rates of Fidelis lead fracture and requirement for re-intervention compared with the other strategies (Table 2).

Figure 3

Sensitivity analysis showing how different values of annual non-device-related mortality affect results for the different strategies. The cumulative percentages of patients experiencing events (inappropriate shocks, re-intervention, and perioperative life-threatening complications) are reported for the first 5 years of follow-up. The analysis shows that the results are robust, i.e. do not depend to a great extent on annual mortality. The upper limit of mortality was set arbitrarily to 20%, as a generator replacement may be considered futile for higher mortalities.

The quality-of-life analyses over a 5-year period are shown in Figure 4. Strategy 1 is clearly the least attractive option, even in the case of a low disutility attributed to inappropriate shocks or requirement for a lead revision, whereas the other strategies were less sensitive to the variations in disutility. The disutility of life-threatening complications did not affect the quality-of-life outcome (the differences were of clinically irrelevant magnitude), due to the low probability of occurrence of these complications.

Figure 4

Quality-adjusted life years (QALYs) at 5 years' follow-up associated with the different strategies resulting from different disutility attributed to inappropriate shocks or re-intervention due to lead dysfunction, and life-threatening complications with lead extraction. Generator replacement only is the worst strategy as from a disutility of inappropriate shocks of ≥0.25 and of re-intervention of ≥0.10. The results are insensitive to life-threatening complications over the entire range of disutility, due to the rarity of these events.


Our decision analysis suggests that systematic revision of Fidelis leads is recommended at generator replacement due to the high probability of lead fracture that will require re-operating over one-third of patients over the following 5 years. A generator replacement without lead revision also exposes the patient to an 11% absolute excess risk of receiving inappropriate shocks due to lead fracture during this time period. Mortality is a major driver of clinical decision-making, but was very similar across the different strategies, as procedural mortality was negligible compared with overall all-cause mortality. Importantly, lead revision remains the preferred option despite the competing risk of overall mortality, and over a wide range of patient perceptions of adverse events such inappropriate shock, requirement for re-intervention, or life-threatening-complications associated with lead extraction.

Three different strategies for Fidelis lead revision were compared. Adding a pacing lead while maintaining the HV component of the Fidelis lead is not recommended in the case of initial Fidelis P/S fracture, due to the high risk of the ensuing HV fracture. However, prophylactic addition of a pacing lead is an option that may be considered, although there is an ∼10% risk of requiring additional lead revision over the next 5–10 years due to subsequent HV fracture, and this may not be the best option in patients at high risk of arrhythmic events. Most physicians are likely to opt for implanting a new ICD lead, with or without extraction of the Fidelis lead. The decision will have to integrate a number of different factors, including patient age and risk profile, operator experience, vein patency, and the number of pre-existing leads (current guidelines22 recommend extraction in case the additional lead results in >4 leads on one side or >5 leads through the superior vena cava). Extraction also carries a risk of unintentionally dislodging or damaging functional leads (e.g. during dissection of fibrous tissue in the pocket to free the ICD lead, or resulting from use of extraction sheaths that may cause collateral damage). This is particularly problematic with coronary sinus leads, which may not be able to be replaced in their original location (e.g. due to thrombosis of the tributary vein).

Our model assumed that the ‘Lead Integrity Alert’ algorithm (which reduces risk of inappropriate shock in case of fracture) was enabled in all patients, but new algorithms and remote monitoring may further reduce this risk. Conversely, patients implanted with generators from manufacturers that do not have this algorithm may be exposed to a higher risk of inappropriate shock. A sensitivity analysis with different rates of inappropriate shocks is presented in the Supplementary material online, the Appendix.

The decision analysis is intended for patients with Fidelis leads, but may be used as a template to construct models for other leads under recall, such as the St Jude Medical Riata lead.

Even though many patients may have died since the lead was originally marketed in September 2004, it was estimated in August 2013 that there are ∼76 000 active patients with Fidelis 6949 leads in the USA alone.24

Study limitations

We did not take into account costs in our model, as these have been addressed in a previous publication.5 Fracture rates were based upon the Fidelis 6949 active-fixation lead, which so far has been the most frequently used model. Recent data25 showed that the Fidelis 6948 passive-fixation lead has a lower failure rate (1.9% per year), which is likely to affect management in these patients. We assumed a constant fracture rate of Fidelis leads (other than in the year of generator replacement), whereas long-term evolution of fracture rates is unclear.2,26 To mitigate this uncertainty, we used published data between 4 and 5 years' follow-up to calculate fracture rates (rather than averaging the rate over the entire 5 years). As with any decision analysis model, the results may not apply to individual patients due to variations in their characteristics. However, the sensitivity analyses that evaluated outcome with different rates of total mortality, lead fracture, inappropriate shock, and procedural complications may help in applying the model to the profile of individual patients. The analysis of Strategy 2 was based upon a 100% lead survival rate for the sake of simplicity. The small incidence of pacing lead fractures encountered in the clinical setting during the timespan of the analysis is however unlikely to significantly affect the results. Finally, our analysis was designed for patients scheduled to undergo a generator replacement, as the dilemma is greatest regarding whether or not to perform lead revision during this intervention. The analysis may not apply for lead revision without a generator replacement, as remaining battery longevity will affect the results.


There is currently no consensus on how to manage patients with functioning Sprint Fidelis leads. Our decision analysis based on published data fills this gap by showing that systematic lead revision (by implanting a new ICD lead rather than a new pacing lead) should be strongly considered during generator replacement to protect the patient from inappropriate shocks and to avoid re-intervention. Whether the Fidelis lead should be extracted or simply abandoned needs to be evaluated on a case-by-case basis.


The work was funded by GECOR (Foundation for Cardiovascular Research of the University Hospital of Geneva). H.B. is funded in part by a research grant from the La Tour Fund for Cardiovascular Research.


The authors wish to thank Prof. Alain Junod for his expert counsel and review of the manuscript.

Author contribution: H.B. prepared the study protocol, performed the data search and extraction, and wrote the manuscript. C.C. verified the source data, performed the statistical analysis, and edited the manuscript. Both authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Conflict of interest: H.B. has received speaker fees and fellowship support from Biotronik, Boston Scientific, Medtronic, Sorin, and St Jude Medical. C.C. has no conflicts of interest to report.


View Abstract