OUP user menu

Lack of current implantable cardioverter defibrillator guidelines application for primary prevention of sudden cardiac death in Latin American patients with heart failure: a cross-sectional study

Jorge Gonzalez-Zuelgaray, Oscar Pellizon, Claudio A. Muratore, Elsa Silva Oropeza, Rafael Rabinovich, José Luis Ramos, Maria Cristina Tentori, Nicolás Reyes, Rubén Aguayo, Jorge Marin, Brett J. Peterson
DOI: http://dx.doi.org/10.1093/europace/eus253 236-242 First published online: 11 September 2012


Aims This cross-sectional study evaluated the application of accepted international implantable cardioverter defibrillator (ICD) guidelines for primary prevention of sudden cardiac death in patients with heart failure.

Methods and results The PLASMA (Probabilidad de Sufrir Muerte Arritmica) study was designed to characterize management of cardiac patients in Latin America. Twelve centres included 1958 consecutively admitted patients in cardiology units in 2008 and 2009. Discharged patients were evaluated for primary prevention, ICD indication and prescription by general cardiologists. Of 1711 discharged patients, 1525 (89%) had data available for evaluating indication status. Class I indications for ICD therapy were met for 153 (10%) patients based on collected data. Only 20 (13%, 95% confidence interval: 7.7–18.4%) patients with indication were prescribed an ICD. Patients prescribed an ICD were younger than patients who were not prescribed an ICD (62 vs. 68 years, P < 0.01). The reasons given by cardiologists for not prescribing an ICD for 133 patients with an indication were: indication criteria not met (75%), life expectancy <1 year (9.7%), rejection by the patient (5.2%), no medical coverage paying for the device (3.7%), psychiatric patient (2.2%), and other reasons (4.2%).

Conclusions In Latin America, international guidelines for primary prevention ICD implantation are not well followed. The main reason is that cardiologists believe that patients do not meet indication criteria, even though study data confirm that criteria are met. This poses a significant challenge and underlines the importance of continuous and improved medical education.

  • Implantable cardiac defibrillators
  • Sudden cardiac death
  • Heart failure
  • Guidelines

What's new?

• The PLASMA study investigated how often ICDs were prescribed by general cardiologists in a series of hospitalized patients in Latin America.

• Only 13% of patients meeting international ICD guideline criteria were prescribed an ICD.

• The main reason for not prescribing an ICD was the general cardiologists' perception that guideline indication criteria were not met (75% of instances), even though collected study data confirmed they were.

• These findings demonstrate a need for increased awareness and committed application of guidelines in Latin American medical practice in order to better serve patients who can benefit from ICD therapy.


Ventricular tachyarrhythmias are the leading cause of sudden cardiac death (SCD) in developed countries. Randomized clinical trials have shown that implantable cardioverter defibrillators (ICDs) reduce mortality from any cause in high-risk populations.1,2 Current international guidelines consider ICDs to be the best therapy choice for patients with ischaemic and non-ischaemic heart disease, heart failure, and ejection fraction (EF) ≤35%.3

The annual incidence of SCD in Latin America is not known. In the general Argentinian population it is similar to what is seen in developed countries.4 However, the annual rate of ICD implants varies considerably between Latin American countries and is typically much lower than in developed countries.5 The lower rate of ICD implants in Latin America may be influenced by several factors including the degree of guideline knowledge, physicians' acceptance of this therapy option, and the availability of resources that allow access to this intervention.

Limited published data exist reporting the number of primary prevention patients in Latin America who meet guidelines and are prescribed an ICD.6 The PLASMA (Spanish acronym for Probabilidad de Sufrir Muerte Arrítmica) study investigated the clinical features and therapy choice for patients in cardiac care centres who met ICD guideline indications. The specific objectives of this study were to (i) estimate how many hospitalized patients meet primary prevention ICD indications in Latin America, (ii) estimate how many indicated patients were prescribed an ICD, (iii) compare the ICD prescription proportion between three well-defined indication groups (based on patient profiles), and (iv) report reasons for not prescribing devices to patients meeting criteria for indication.


PLASMA was a cross-sectional study carried out in 12 Latin American medical centres (Argentina, Chile, Colombia, Mexico, and Dominican Republic) between August, 2008 and December, 2009. Electro-physiologists served as the principal investigator at each centre. The study protocol was approved by the ethics committees at each participating centre and all patient participants provided informed consent. This study complies with the Declaration of Helsinki. The Appendix shows a list of all participating centres and researchers.


All patients hospitalized for any cause at a coronary care unit or cardiology service were eligible to participate. To avoid possible selection bias, all hospitalized patients were included except for those who died during hospitalization because they would have no chance to have an ICD prescribed. The number of participants was solely based on how many patients were hospitalized during the study timeframe.

Demographic data, personal history, reason for admission, physical examination, complementary examinations (electrocardiogram, echocardiogram, coronary angiography, electrophysiological study, among others), in-hospital evolution, and treatment at discharge were recorded on a prescribed form during the hospitalization. No post-hospitalization follow-up data were collected.


During hospitalization, the treating physician (general cardiologist) determined whether the patient was an ICD candidate (including cardiac resynchronization devices with defibrillators) for primary prevention of sudden death using their knowledge of the guidelines, professional judgement, and patient information collected on the study form (class I in ACC/AHA/HRS 2008 guidelines).3 Patients with an acute myocardial infarction (AMI) within the previous 40 days, those who underwent heart surgery or catheter revascularization for coronary disease in the previous 3 months, or patients planning on undergoing revascularization or valve surgery were excluded as candidates for implantation. If the patient was determined to be an ICD candidate, the treating physician, referring general physician, and patient would discuss and decide if ICD therapy would be used. Whether or not an ICD was prescribed and the reasons for not prescribing were documented.

All discharged patients were classified as meeting ICD indications or not according to the ACC/AHA/HRS guidelines using collected study data. Patients without enough data to make a classification were termed as ‘unknown if indicated’ and excluded from the main analysis. Class I indications fell into three groups. In Group 1, an ICD is indicated by the presence of an EF ≤ 35% due to the previous MI (>40 days) and New York Heart Association (NYHA) heart failure functional class II or III (level of evidence A). In Group 2, an ICD is indicated by the presence of non-ischaemic dilated cardiomyopathy, EF ≤ 35%, and NYHA heart failure functional class II or III (level of evidence B). In Group 3 an ICD is indicated by the presence of a previous MI (>40 days), EF ≤ 30%, and NYHA heart failure functional class I (level of evidence A). The ICD prescription proportion for patients meeting indications was calculated overall and for the three groups.

Statistical methods

For reporting patient characteristics, proportions were used for categorical variables and means with standard deviations were used for continuous variables. Comparisons between patients with ICDs prescribed and not prescribed were performed using the Wilcoxon–Mann–Whitney test for numerical or ordinal variables and the χ2 test or Fisher's exact test for categorical variables. The χ2 test was used to compare indication groups. Two-tailed P values were reported. Statistical significance was claimed for values of P < 0.05. Confidence intervals (CIs) for proportions were constructed using the normal approximation method.



One thousand, nine hundred and fifty-eight patients were enrolled (1125 men and 833 women) with a mean age of 64 ± 14.6 years (range 7–99). Patients who died during hospitalization were not included (n = 247), leaving 1711 patients with collected data. Patient diagnosis at admission was acute coronary syndrome without ST-segment elevation in 36% (n = 611), heart failure in 15% (n = 247), cardiac arrhythmia in 26% (n = 453), AMI in 14% (n = 246), and ‘other’ in 9% (n = 154). Information for determining indication group was documented for 1525 (89%) of the 1711 discharged patients. Patients classified as ‘unknown if indicated’ differed by diagnosis at admission (9% for acute coronary syndrome, 17% for heart failure, 16% for cardiac arrhythmia, 0% for AMI, and 11% for other, P < 0.001). After excluding patients with MI <40 days (n = 246), cardiac intervention (n = 245), no heart failure (n = 547), NYHA class IV (n = 18), EF > 35% (n = 316), and those ‘unknown if indicated’ (n = 186), 153 patients with an ICD indication remained for further analyses. Figure 1 depicts the complete patient flow. Of all 1525 discharged patients with a known indication status, 153 (10%) met a primary prevention indication according to ICD therapy guidelines.

Figure 1

Flow diagram of patient selection for analysis. All hospitalized patients from 12 Latin American medical centres between August 2008 and December 2009 were considered for analysis. Patients for which indication status could not be determined due to missing data were classified as ‘unknown if indicated.’ AMI, acute myocardial infarction; EF, ejection fraction; HF, heart failure; NYHA, New York Heart Association; PTCA, percutaneous transluminal coronary angioplasty.

One hundred and fifty-three patients with implantable cardioverter defibrillator indications

Clinical characteristics of the patients who met ICD indications are shown in Table 1. The patients were predominantly male (75.8%) with a mean age of 67.1 ± 13 years (range 32–94). Heart failure aetiologies were reported as ischaemic (61.3%), idiopathic (8.1%), valvular (10.5%), hypertensive (5.7%), Chagasic (5.7%), alcoholic (4%), and chronic obstructive pulmonary disease (1.6%). The pharmacological treatments were angiotensin enzyme inhibitors or angiotensin II receptor blockers (78%), beta-adrenergic blockers (75%), aldosterone antagonists (46%), and amiodarone (25%). Ninety-eight patients were in Group 1 (64%), 38 in Group 2 (25%), and 17 in Group 3 (11%). Fifty-one patients (33%) met MADIT II criteria for primary prevention of sudden death, 102 (67%) met SCD-HeFT criteria, and no patients met MADIT I criteria.

View this table:
Table 1

Clinical characteristics for patients meeting implantable cardioverter defibrillator indications

VariableALL (n = 153)ICD (n = 20)No ICD (n = 133)P value
Age (years) (mean, SD)67.1 ± 13.061.5 ± 9.668.0 ± 13.30.012
Male116 (76%)17 (85%)99 (74%)0.41
Diabetes48 (31%)6 (30%)42 (32%)0.89
Dyslipidaemia53 (35%)6 (30%)47 (35%)0.64
Obesity30 (20%)8 (40%)22 (17%)0.029
Smoking35 (23%)8 (40%)27 (20%)0.08
Hypertension109 (71%)15 (75%)94 (71%)0.69
Alcoholism10 (67%)1 (5%)9 (7%)1.00
Myocardial infarction61 (40%)10 (50%)51 (38%)0.32
NYHA III38 (25%)8 (40%)30 (23%)0.10
Atrial fibrillation51 (34%)4 (20%)47 (35%)0.17
LVEF (mean%, SD)26.3 ± 5.925.3 ± 5.626.5 ± 6.30.42
QRS duration (ms) (mean, SD)112 ± 29129 ± 31109 ± 280.012
Idiopathic10 (7%)2 (10%)8 (6%)0.62
Ischaemic75 (49%)11 (55%)64 (48%)0.57
Chagas7 (5%)3 (15%)4 (3%)0.048
ACEI/ARB119 (78%)15 (75%)104 (78%)0.78
Beta blockers115 (75%)13 (65%)102 (77%)0.27
Spironolactone71 (46%)14 (70%)57 (43%)0.023
Amiodarone38 (25%)9 (45%)29 (22%)0.048
  • ARB, angiotensin receptor blockers; ACEI, angiotensin-converting enzyme inhibitors; CM, cardiomyopathy; ICD, implantable cardioverter defibrillator; LVEF, left ventricle ejection fraction; NYHA, New York Heart Association.

Implantable cardioverter defibrillator prescription

Of the 153 patients with a class I ICD indication according to the guidelines, 20 (13.1%, 95% CI: 7.7–18.4%) had an ICD prescribed [includes cardiac resynchronization therapy-defibrillator (CRT-D) prescriptions]: 17 during hospitalization and 3 prior to hospitalization. Four patients were prescribed a CRT-D and 16 an ICD. Table 1 shows the clinical characteristics for patients who were prescribed an ICD and those who were not. Patients prescribed an ICD were younger than patients who were not (62 vs. 68 years, P < 0.01), had more Chagas' disease, were more often obese and were more likely to use spironolactone and amiodarone. An ICD was prescribed in 13.2% of Group 1 patients (13/98), 18.4% of Group 2 patients (7/38), and 0% of Group 3 patients (0/17; Figure 2). The proportion of ICD prescriptions did not differ significantly between the three groups (P = 0.17). None of the 186 patients who were ‘unknown if indicated’ had an ICD prescribed.

Figure 2

ICD prescription overall and by indication group. Number and percent of implantable cardioverter defibrillator prescriptions for primary prevention patients indicated by ACC/AHA/HRS 2008 guidelines overall and by indication group. Group 1: implantable cardioverter defibrillator indicated in the presence of an ejection fraction ≤35% due to previous myocardial infarction (>40 days) and New York Heart Association functional class II/III heart failure. Group 2: implantable cardioverter defibrillator indicated in the presence of non-ischaemic dilated cardiomyopathy, ejection fraction ≤35%, and New York Heart Association functional class II/III heart failure. Group 3: implantable cardioverter defibrillator indicated in the presence of previous myocardial infarction (>40 days), ejection fraction ≤30%, and New York Heart Association functional class I heart failure.

The main reasons general cardiologists gave for not prescribing an ICD for 133 patients with indication (according to study data) were ‘indication criteria not met’ (75%), life expectancy <1 year (9.7%), rejection by the patient (5.2%), no medical coverage paying for the device (3.7%), psychiatric patient (2.2%), and other reasons (4.2%). When considering the 34 patients for whom general cardiologists believed indication criteria were met, the distribution of reasons became life expectancy <1 year (38.2%), rejection by the patient (20.5%), no medical coverage paying for the device (14.7%), psychiatric patient (8.8%), and other reasons (17.6%; Figure 3). ‘Indication criteria not met’ was the most frequent reason cited by the physician in all three indication groups, with reported frequencies of 68 (80%), 16 (52%), and 15 (88%) in Groups 1, 2, and 3, respectively. General cardiologists were more likely to report this reason for patients in Groups 1 and 3 compared with Group 2 (P = 0.009).

Figure 3

Main reasons given by general cardiologists for not prescribing an ICD to patients meeting guideline indications according to study data. Left panel: includes all 153 patients meeting guideline indications according to study data. ‘Criteria not met’ was the dominant reason (75%). This means the general cardiologist did not think implantable cardioverter defibrillator guideline indication criteria were met even though the collected study data confirm that they were. Right panel: displays the distribution of reasons for not prescribing an implantable cardioverter defibrillator to the 34 patients (of all 153) for whom the general cardiologist believed indication criteria were met and the collected study data confirm they were (i.e. excludes patients in the left panel with ‘criteria not met’).


Patients with an indication and implantable cardioverter defibrillator prescriptions

PLASMA provides insights into real-world practices for primary prevention of SCD in Latin America. This cross-sectional sample showed that 10% of patients hospitalized for any reason at cardiology services met guideline indications for ICD therapy. Thirteen per cent of patients meeting indication criteria were prescribed an ICD. The ICD prescription proportion did not differ between three defined indication groups based on patient measurements of EF, NYHA class, ischaemic aetiology, and history of MI. The observed ICD prescription rate in PLASMA is lower than reported in the GWTG-HF registry7 in which only one of every five potentially eligible patients either arrived at implant or were told about ICD therapy at discharge. The IMPROVE HF study reported that 51% of eligible US patients received ICD or CRT-D therapy.8 Other studies have found a low adherence to guidelines for ICD implantation,912 although this appears to be even lower in Latin America.

Reasons for not prescribing an implantable cardioverter defibrillator for patients with criteria for indication

The most cited reason (75% overall) for not prescribing an ICD during hospitalization was ‘indication criteria not met’ in all three patient groups, even though these patients met indications according to collected study data. The most likely explanation is that general cardiologists in Latin America, often not well versed in device therapy, were not aware of or meticulously applying the guidelines criteria. A previous study showed that the lack of electro-physiologists or limitations in hospital facilities were significant causes for low implant rates.7 In PLASMA, participating centres had state-of-the-art technology and personnel specializing in ICD implants and management, including electro-physiologists. However, when general cardiologists care for patients in Latin American coronary care units, typical practice does not usually include consulting an electro-physiologist.

It is also possible that other patient characteristics were considered in addition to guidelines criteria. In this population, patients prescribed an ICD were younger and more likely to use spironolactone and amiodarone compared with those not prescribed an ICD. Voller et al.11 also reported that younger patients were more likely to receive an ICD.11 PLASMA patients meeting primary prevention indications were 67 years old on average, which is higher than SCDHeFT (60 years) and MADIT II (64 years).1,2 However, advanced age is not a contraindication for ICD therapy.13 Another explanation is that physicians were content to use medications first even though ICDs have proven survival benefit compared with conventional medical therapy.1,2 Several randomized trials have shown that angiotensin-converting enzyme inhibitors,1416 beta blockers,17,18 and aldosterone antagonists19 improve survival in patients with heart failure. The percentage of patients receiving these drugs at discharge was 78, 75, and 46%, respectively, which is comparable to reports from developed countries.2025 Amiodarone was prescribed for 25% of patients which is typical practice in Latin America for treating atrial and ventricular arrhythmias. In our study, 34% of patients had atrial fibrillation for which the first drug treatment is amiodarone.

Economic considerations

The cost of treatment is another decision-making point in ICD prescription. Cost-effectiveness studies conducted in the USA and Europe showed an acceptable result for ICD therapy.26,27 Alcaraz et al.28 studied an Argentinian population, concluding that devices are clearly cost-effective in the MADIT I population and are moderately cost-effective for patients with MADIT II characteristics or in secondary prevention. Mark et al.29 found ICDs to be economically efficient at improving health benefits in SCD-HeFT patients. In PLASMA, 33% of patients met MADIT II criteria and 67% met SCD-HeFT criteria for primary prevention of SCD. Ribeiro et al.30 showed that the costs associated with ICD treatment are proportionately higher in Brazil compared with developed countries, but the cost-benefit ratio still favours the ICD. Given that 14.9% of patients meeting criteria for indication in the PLASMA study did not receive an ICD due to a financial reason, cost played a modest role in not prescribing ICD therapy.

Patient refusal

Current medicine is faced with a paradigm change in which patients have become active participants in the decision-making process regarding their health. In PLASMA, when physicians believed ICD indications were met and discussed that therapy option, 20.9% of patients refused. Voller et al.11 reported 29.2% of patients denied ICD therapy; a sizeable impact on ICD therapy application. The ‘availability heuristic’ of Tversky and Kahneman31 shows that behaviour is influenced by how easily an event is remembered. Thus, the decision to use a therapy is affected by physician and patient knowledge of the risk-benefit ratio, the way in which the information is presented and past experiences.


The guidelines recommend ICD therapy without distinction between outpatients or inpatients. All patients included in this study were hospital inpatients at state-of-the-art medical institutions. Thus the indication and ICD prescription rates reported may not represent the broader Latin American population. One hundred eighty-six patients (10.8%) were classified as ‘unknown if indicated’ thereby excluding them from analysis. The ‘unknown’ classification differed by admission reason, most notably being 0% for AMI admissions because they were automatically classified as not meeting indication criteria (MI < 40 days). No patients with an ‘unknown’ classification were prescribed an ICD. However, how many of those patients would have met indications if pertinent information was available remains unknown.


The PLASMA study investigated patients hospitalized in Latin American cardiac care centres and found that only 13% who met the ACC/AHA/HRS guidelines class I criteria for primary prevention ICD indication had a device prescribed. The main reason cited by general cardiologists for not prescribing an ICD was ‘indication criteria not met’ (75% of patients), even though study data confirmed indication. This reveals a lack of adherence to internationally accepted guidelines and emphasizes the need for increased awareness and committed application of guidelines in Latin American medical practice. This may be achieved through increased collaboration with electro-physiologists and more effective and continuous medical education by universities, medical institutions, and scientific societies.

Faced with the dual epidemic of SCD and heart failure, increasing application of cost-effective ICD therapy could save thousands of lives.

Conflict of interest: C.A.M. and B.J.P. are employed by Medtronic.


The study was supported by Medtronic Inc. Latin America Operations.


PLASMA investigators

Jorge González Zuelgaray, Ariel Szyszko, Sanatorio de la Trinidad San Isidro, Argentina. Jose L. Ramos, Laura Sanziani, Hospital Italiano de Rosario, Argentina. Oscar Pellizzon, Antonia Catalano, Sebastián Nannini, Hospital Provincial del Centenario, Rosario, Argentina. Rafael Rabinovich, Sanatorio Modelo de Quilmes, Argentina. María C Tentori, Sandra Cepeda, Hospital Fernandez, Argentina. Elsa Silva Oropeza, Horacio Rodríguez Pina, Centro Médico Siglo XXI IMSS, México. Nicolás Reyes Reyes Hospital Ángeles de Puebla, México. Santiago Nava, María Garcia Merodio, Moisés Levinstein, Hospital ABC, México. Milton Guevara Valdivia, Hospital CMN La Raza, México. Rubén Aguayo, Hospital San Juan de Dios, Santiago, Chile. Jorge Marín, Clínica Medellín, Colombia. Pedro Ureña, Valerie Izarnotegui, CEDIMAT, Republica Dominicana.


View Abstract