Use of an entacapone-containing drug combination and risk of death: Analysis of the FDA AERS (FAERS) database

Abstract

To assess the signal of death associated with the use of an entacapone-containing drug combination in the FDA Adverse Event Reporting System (FAERS) database.

Reports of death events submitted between January 2004 and December 2010 were retrieved and analysed by the reporting odds ratio (ROR). The ROR of case/non-case reports of death associated with an entacapone-containing drug combination was compared with the levodopa/carbidopa combination using the FDA AERS database.

Eighty-seven reports linked the entacapone-containing drug combination to death, compared to 27 reports of death linking the levodopa/carbidopa combination. The ROR was statistically significant for the association between deaths with the use of an entacapone-containing drug combination (1.86 [95% CI 1.50–2.31]). In contrast, the ROR of death associated with the combination of levodopa and carbidopa was not statistically significant (0.89 [95% CI 0.61–1.30)].

Based on analysing reports in the FAERS database, there is a risk of death with the use of an entacapone-containing drug combination. These results generated a signal of death with the use of this drug. However, epidemiological studies are required to confirm this association.

1. Introduction

Entacapone is a catecholamine-o-methyl transferase (COMT) inhibitor. The drug was approved in October 1999 as a single ingredient (marketed as Comtan®) and in June 2003 as a combination pill containing levodopa, carbidopa and entacapone (marketed as Stalevo®). It is indicated as an adjunct to levodopa/carbidopa therapy in patients with Parkinson’s disease (PD) who experience the signs and symptoms of end-of-dose “wearing-off”. It has been found that up to 50% of patients develop motor complications and end-of-dose “wearing-off” after 5 years of treatment with levodopa (Standaert and Young, 2001). COMT inhibitors are added to levodopa treatment to overcome these complications and to prolong the bioavailability of levodopa (Rivest et al., 1999).

In recent years, concerns have been raised about the safety of COMT inhibitors. Tolcapone was the first COMT inhibitor to be approved, but it was withdrawn in the European Union and in other countries after rare reports of death due to hepatic toxicity (World Health Organisation, 2001; Olanow and Watkins, 2007). On the other hand, entacapone was considered to be safe and well-tolerated with regard to hepatotoxicity (Brooks, 2004).

However, recent findings from the Stalevo Reduction in Dyskinesia Evaluation-Parkinson Disease (STRIDE-PD) study have raised concerns over the safety of entacapone (Brooks, 2004; Stocchi et al., 2010). The STRIDE-PD study compared levodopa/carbidopa with or without entacapone in 747 patients with PD. The trial reported an imbalance in some serious outcomes, including the incidence of prostate cancer and myocardial infarction in patients treated with levodopa, carbidopa and entacapone compared to those receiving only levodopa/carbidopa (Stocchi et al., 2010; Food and Drug Administration, 2010a).

On March 31 2010, the US Food and Drug Administration (FDA) announced that they were evaluating data from the STRIDE-PD clinical trial which suggested that patients taking Stalevo® may be at an increased risk of developing prostate cancer (Food and Drug Administration, 2010a). The FDA has not yet concluded that Stalevo® increases the risk of developing prostate cancer. Moreover, on August 20, 2010, the FDA conducted a meta-analysis that included 15 clinical trials based on the findings from the STRIDE-PD trial. A small increased risk of cardiovascular adverse events was found in the Stalevo® group.(Food and Drug Administration, 2010b).

The purpose of this study was to assess the signal of death associated with the use of an entacapone-containing drug combination based on the cases of death submitted to the FDA Adverse Events Reporting System (FAERS) database.

2. Methods

The FDA adverse event reporting system (FAERS) was utilised to assess the association between the use of an entacapone-containing drug combination and the signal of death using a case/non-case method. All reports that were available in the FDA AERS database from January 2004 to December 2010 were accessed and downloaded from the FDA website (http://www.fda.gov/). However, in the dataset from 2004, there were some reports dated before 2004. FAERS is a spontaneous reporting system that contains data on adverse drug events and medication errors submitted to the FDA (Food and Drug Administration, 0000).

The FDA AERS contains either all or some of the information on the MedWatch reporting form that is usually sent by healthcare professionals as well as patients or consumers from the United States (U.S.) or outside the U.S. In addition, this system includes reports sent from the manufacturers for serious adverse drug reactions in the U.S. or serious and unlabelled spontaneous reports from outside the U.S. It is worth noting that any report should include four elements to be considered a valid report prior to entry into the system: an identifiable patient, an identifiable reporter, the suspected drug and the adverse drug event. The reporter and suspected drug elements are available as free text while the adverse drug events are coded by the Medical Dictionary for Regulatory Activities (MedDRA). The MedDRA term level used to code the adverse drug event is the preferred term (PT) Food and Drug Administration, 0000; Poluzzi et al., 2009, 2010.

The main group of medications of interest was drugs that include entacapone for the treatment of Parkinson’s disease. Stalevo® is the only drug that contains entacapone (Food and Drug Administration, 2010a). Another group of medications was also included in this study, i.e. the combination of levodopa and carbidopa, which is marketed as Sinemet®. Entacapone is either used as a separate pill with levodopa and carbidopa or as a combination of the three drugs in one pill, as in the case of Stalevo® (Food and Drug Administration, 2010a). Both the trade and generic (chemical name) names were used to extract the reports that included entacapone-containing products as well as other drugs included in this study. The outcome of interest was death, and was searched using the term “death” in the PT field in the FAERS database.

A case/non-case method was used to analyse the association between the use of drugs of interest and the signal of death. Cases were the reports of the outcome of interest, which was “death”, and were extracted using the PT for any given drug, while the non-cases were all other reports related to the same drug (other than the reports of the outcome of interest, which was death). All death terms that were not related to the population of the study, for example “neonatal death” and “inter-uterine death”, were excluded. For both cases and non-cases, only reports that considered the drug of interest as the primary suspected drug were included in the analysis. The reporting odds ratio (ROR), as a measure of disproportionality, was used to assess the association between the drug of interest and death. The ratio of case/non-case reports for each group (the entacapone-containing drug combination group and the levodopa/carbidopa group) were compared to other medications. The ROR calculation is explained in detail elsewhere (Moore et al., 1997). All statistical analyses were performed using SAS® software (version 9.2) (SAS Institute Inc., Cary, North Carolina, USA).

3. Results

From the period between January 2004 and December 2010, there were 5062 reports that included entacapone; of these reports, there were 2532 reports in which the drug was considered the primary suspected cause of the outcome of interest (all outcomes). Of these reports, there were 87 reports that linked the entacapone-containing drug combination to be the primary suspected cause of death. For medications including levodopa and carbidopa only, out of 9477 reports, there were 1670 reports that considered the drug to be the primary suspected cause of the outcome of interest. Of these reports, there were 27 reports that linked the levodopa and carbidopa combination to the primary suspected cause of death.

The ROR of death associated with the entacapone-containing drug combination was statistically significant (1.86 [95% CI 1.50–2.31]). Conversely, the ROR of death associated with the combination of levodopa and carbidopa was not statistically significant (0.89 [95% CI 0.61–1.30)] (Table 1). Age was recorded in 48 cases of those who were using the entacapone-containing drug combination; the mean age was 75 years with a minimum age of 50 years and a maximum age of 93 years. Only seven cases of patients who were using the combination of levodopa and carbidopa contained the age information; the mean age was 72, the minimum age was 40 years and the maximum age was 87 years. The majority of cases in both groups were males (entacapone group 64% and levodopa/carbidopa group 73%).

Table 1.

ROR of death for the levodopa/carbidopa/entacapone and levodopa/carbidopa drug combinations.

Active substances	Death events	All ADRs	ROR	95% CI
Levodopa/carbidopa/entacapone	87	2532	1.86	1.50–2.31
Levodopa/carbidopa	27	1670	0.89	0.61–1.30

ADR: adverse drug reaction, ROR: reporting odds ratio, 95% CI: 95% confidence interval.

A stratified year-by-year analysis was performed for events of death occurred in levodopa/carbidopa/entacapone group and levodopa/carbidopa group (Fig. 1). We noticed that the peak of cases for the entacapone group was in 2009 and 2010, while the peak of cases for levodopa and carbidopa was in 2006 and 2007. Five cases in the entacapone group occurred in clinical studies.

Figure 1.

Number of events of death in levodopa/carbidopa/entacapone group and levodopa/carbidopa group stratified by year.

In the entacapone-containing drug combination group, although all cases had information on the primary suspected drug, some cases had information on concomitant drug use. As concomitant medications, other anti-Parkinson’s disease drugs (e.g. pergolide and pramipexole), hypotensive agents (e.g. angiotensin converting enzyme inhibitors) and antidepressants (e.g. selective sertonin reuptake inhibitors) were the concomitant medications in the entacapone-containing drug combination group. In contrast, few patients had information on concomitant drug use in the levodopa/carbidopa group. Examples of these drugs include anti-Parkinson’s disease drugs (e.g. pramipexole), lipid lowering agents (e.g. statins) and antidepressant agents (selective sertonin reuptake inhibitors).

4. Discussions

The Food and Drug Administration Adverse Event Reporting System (FAERS) is one of the largest databases (it has over four million reports at the current time) and is considered to be a source for data mining and signal detection in addition to the World Health Organisation database (VigiBase™) that includes more than seven million reports (Poluzzi et al., 2009; Food and Drug Administration, 0000; World Health Organisation, 0000). However, the FAERS database can be freely accessed without charge to allow researchers to investigate any signal from these reports (Poluzzi et al., 2009). The FAERS database is a very useful and important tool for post-marketing surveillance and for pharmacovigilance and drug safety activities, especially for serious events (Hochberg et al., 2007).

In this study, we used a case/non-case method, which has been utilised by many studies (Poluzzi et al., 2009; Piccinni et al., 2011; Motola et al., 2012). This method has been explained in detail elsewhere (Moore et al., 1997). Our study found an association between the use of the entacapone-containing drug combination and death. While this association was not seen in patients using the levodopa/carbidopa combination, the risk would be the same since there is only one extra substance in Stalevo® (levodopa, carbidopa and entacapone) compared to Sinemet® (levodopa and carbidopa). However, Stalevo® (levodopa/carbidopa/entacapone) has several safety concerns, especially following the STRIDE-PD study. In March 2010, the FDA published a safety communication on the risk of prostate cancer with Stalevo® use. The FDA was evaluating the STRIDE-PD study because this study raised concerns over the safety of entacapone. The STRIDE-PD study compared levodopa/carbidopa with or without entacapone in 747 patients with PD (Stocchi et al., 2010). The trial reported an imbalance in some serious outcomes, including the incidence of prostate cancer in patients treated with levodopa, carbidopa and entacapone compared to those receiving only levodopa/carbidopa. There were nine cases of prostate cancer in the 245 participants (3.7%) in the Stalevo® group, while there were two cases out of 222 participants (0.9%) in the levodopa/carbidopa group. The incidence of prostate cancer was 14 per 1000 patient years for the Stalevo® group and 3.2 per 1000 patient years for the levodopa/carbidopa group (Stocchi et al., 2010; Food and Drug Administration, 2010a). However, other trials before the STRIDE-PD study did not raise any safety concerns as this trial. Yet, most of these trials were conducted over a short period of time (less than a year) compared to the STRIDE-PD study, which was conducted for over 4 years (Food and Drug Administration, 2010a).

Furthermore, on August 2010, the FDA also announced a safety communication mentioning that there is an increased risk of cardiovascular events, such as heart attack, stroke and cardiovascular death with the use of Stalevo® compared to patients using levodopa/carbidopa based on data from a clinical trial. The risk of cardiovascular events was first found in the STRIDE-PD study, as there were seven cases of myocardial infarction and one case of cardiovascular death in the Stalevo® group while there were zero cases of both myocardial infarction and cardiovascular death in the levodopa/carbidopa group. Moreover, the FDA performed a meta-analysis to investigate this risk. This meta-analysis included data from 15 clinical trials. Twenty-seven cardiovascular events occurred in the Stalevo® group and 10 events occurred in the levodopa/carbidopa group; the relative risk was 2.46, 95% CI 1.19, and 5.09. However, the method of performing this meta-analysis has several limitations (Food and Drug Administration, 2010b). Cardiovascular death was one of the endpoints investigated in both the STRIDE-PD study and the meta-analysis performed by the FDA. Therefore, we used death (all causes of death) as our endpoint in this study. The results of our study are similar to the results of the meta-analysis performed by the FDA and consistent with the results of the STRIDE-PD trial with respect to the occurrence of death (Food and Drug Administration, 2010b).

Parkinson’s disease is not a fatal disease and is not associated with high mortality. However, the risk increased in patients who developed dementia (de Lau et al., 2005). Therefore, we believe that the risk of death associated with patients in the levodopa/carbidopa/entacapone group is not related to the disease itself, because the risk only occurred in the levodopa/carbidopa/entacapone group while it did not occur in the levodopa/carbidopa group, despite the fact that both groups had the same disease. Concomitant medications were used as a proxy for concomitant disease to assess the risk factors for those patients. Most of these drugs were used for common diseases that affect this age group, such as hypertension and hyperlipidaemia. Furthermore, some of the patients in both groups had depression. It does not seem that these risk factors had a major effect on the study outcome. Although antidepressant use is associated with the risk of suicide, this does not seem important here because (1) both groups were using antidepressants and (2) the risk of death was higher only in patients in the levodopa/carbidopa/entacapone group.

Notoriety bias occurs when the number of reports increases following safety alerts, especially those from regulatory bodies (Tuccori et al., 2010). Although notoriety bias could have had an effect on the results of our study, we do not believe that the association between Stalevo® and death can be explained only by notoriety bias. We noted that there was an increasing trend in the risk of death from 2004 to 2009. Furthermore, the number of cases of death in 2010 was less than in 2009 in the entacapone group. The demographics of the cases of death in our study were consistent with the epidemiology of Parkinson’s disease. The mean age was more than 60 years and the risk was more in males than females (de Lau and Breteler, 2006). To our knowledge, the present study is the first study to use the FAERS database to investigate the signal of death with Stalevo® and levodopa/carbidopa.

As mentioned previously, this study has several advantages such as using rich open access data to assess medication safety signals. This database includes millions of reports and covers the entire American population and also contains reports from other countries. This advantage is important to investigate rare events. This advantage allows for an independent external researcher to evaluate these medications and is not limited to investigations by US FDA staff. Also, this is the first study to use the FAERS database to assess the risk of death with the use of commonly used anti-Parkinson’s medications.

However, this study has several limitations, especially with respect to the case/non-case method and any spontaneous reporting system, i.e. (1) under-reporting bias, (2) over-reporting bias which happens due to recent safety alerts and the notoriety bias, (3) missing of data or lack of information, especially regarding patient lifestyle information or occupational risks, (4) the Weber effect, which is an increase in the number of reports following the approval of new drugs (Hartnell et al., 2003) and (5) the inability to assess the drug-event causal relationship. These limitations are common with all studies that use spontaneous reporting systems and ROR analysis.

5. Conclusion

The results of our study found an association between the use of the entacapone-containing drug combination and death; this association was not seen in patients using the levodopa/carbidopa combination.

Despite the limitations of the case/non-case method and the FDA AERS database, we believe that the data in this study may offer an important reference for defining the safety profile of entacapone. Further epidemiological studies (case control or cohort studies) should be utilised to investigate the safety of this medication.

Disclosure

Authors have no conflict of interests to declare.