Do antiepileptic drugs or generalized tonic–clonic seizure frequency increase SUDEP risk? A combined analysis

Summary

Purpose:  In an analysis of four case–control studies of sudden unexpected death in epilepsy (SUDEP), we found that yearly frequency of generalized tonic–clonic seizures (GTCS) and antiepileptic drug (AED) polytherapy were associated with an increased risk for SUDEP. The prior analysis, however, did not evaluate AEDs and GTCS frequency concurrently.

Methods:  We combined data from the three case–control studies with information on the frequency of GTCS and AED therapy, that is, carbamazepine, phenytoin, valproic acid, and other AED therapy. Number of AEDs was also considered. Lamotrigine and GTCS frequency were considered separately in two of the case–control studies. Logistic regression analysis was used to evaluate GTCS frequency, each of the AEDs, and number of AEDs. Adjusted analysis of the different AEDs accounted for study, age at death, gender, and GTCS frequency.

Key Findings:  In crude analysis, GTCS frequency, AED polytherapy, and number of AEDs were associated with an increased risk for SUDEP. Analysis of individual AEDs and of number of AEDs, adjusting for GTCS frequency, revealed no increased risk associated with AEDs as monotherapy, polytherapy, or total number. GTCS frequency remained strongly associated with an increased risk for SUDEP.

Significance:  Our findings—that none of the AEDs considered were associated with increased SUDEP risk as monotherapy or as polytherapy when GTCS frequency was taken into account—provide a consistent message that number of GTCS increases SUDEP risk and not AEDs. These results suggest that prevention of SUDEP must involve increased efforts to decrease GTCS frequency in order to avert the occurrence of this devastating epilepsy outcome.

In a combined analysis of the four case–control studies of SUDEP with living epilepsy controls (Hesdorffer et al., 2011), we found that the risk for SUDEP was increased by number of generalized tonic–clonic seizures (GTCS) each year and by antiepileptic drug (AED) polytherapy; lamotrigine therapy was also associated with an increased risk for SUDEP. Our prior analysis did not simultaneously evaluate AEDs and GTCS frequency and thus we were unable to consider whether the increased risk for SUDEP was associated with GTCS frequency, AED polytherapy, or both, or with any specific AED. Here, we examine the association between SUDEP and each of the AEDs that were taken by a sufficient number of cases and controls to permit analysis while adjusting for GTCS frequency.

Methods

We combined the data from three of the four case–control studies of risk factors for SUDEP. These studies had information on GTCS frequency per year and AED therapy (Nilsson et al., 1999a; Walczak et al., 2001; Langan et al., 2005). SUDEP was defined as (1) a history of epilepsy (more than one epileptic seizure during a period of <5 years); (2) death occurring suddenly; (3) death unexpected (i.e., there was no life threatening illness); and (4) death remaining unexplained after all investigative efforts, including autopsy. Definite SUDEP required autopsy, whereas probable required criteria 1–4, except autopsy. Controls were living people with epilepsy. Excluded were cases and controls with a history of heart disease. The methods of each of these studies have been described previously (Hesdorffer et al., 2011).

Factors were defined according to the categorization that could be applied to all three data sets. AEDs separately considered in the three studies were carbamazepine, phenytoin, valproic acid, and other AEDs. These were categorized as monotherapy, polytherapy, no AED, or other AED. The number of AEDs was separately analyzed both with four or more and separately as three or more as the highest category. Additional analyses were conducted in the subset of the two studies (Walczak et al., 2001; Langan et al., 2005); these included information on lamotrigine therapy. Potential confounders included age at death, gender, study (data source), and yearly frequency of GTCS categorized as 0, 1–2, >3, or unknown.

Data were analyzed with SAS (SAS Institute, Cary, NC, U.S.A.), using the Mantel-Haenszel method for categorical variables with testing at the two-tailed level of 0.05. Excluded from the analysis were 17 subjects missing AED information and five subjects missing age at death. Analyses were conducted using logistic regression, ignoring the matching present in the individual studies. Crude analysis was conducted for each AED and for GTCS frequency adjusted for study with the English study as the referent. Adjusted analysis of the different AEDs accounted for study, age at death, gender, and GTCS frequency per year.

Results

There were 216 cases and 831 controls in the analysis of carbamazepine, phenytoin, valproic acid, other AED, and GTCS frequency (Table S1). The English study was the source of most cases and controls. More than half of the cases and controls were younger than16 years of age at epilepsy onset, and males predominated.

In the analysis of lamotrigine and GTCS frequency, there were 160 cases and 674 controls (Table S2). Two-thirds of the cases and controls had an age of onset of epilepsy of younger than16 years, and more than half were male.

Yearly GTCS frequency was associated with an increased risk for SUDEP in crude analysis (Table 1). Similarly, polytherapy was associated with an increased SUDEP risk for carbamazepine, phenytoin, valproic acid, and other AED therapy in crude analysis. After adjustment for GTCS, age at death, gender, and study, none of the AEDs were associated with an increased SUDEP risk, but number of GTCS per year remained strongly associated with an increased risk for SUDEP. Number of AEDs was no longer associated with an increased risk for SUDEP after adjustment for GTCS, age at death, gender, and study. This was consistent whether four or more AEDs or three or more AEDs were the highest category. Compared with no treatment, AED monotherapy appeared protective, although the upper limit of the 95% confidence interval (95% CI) was close to 1, odds ratio (OR) 0.5 (95% CI, 0.3–0.995) (Table 1).

Table 1.   Odds ratios and 95% confidence intervals for antiepileptic drugs, primary or secondary generalized tonic–clonic seizure (GTCS) frequency, and risk for SUDEP data from England, Sweden, and the United States

Factors	Crude analysis^b	Adjusted analysis^a
		GTCS	CBZ and GTCS	PHT and GTCS	VPA and GTCS	Other and GTCS	Number of drugs and GTCS
N = 1,042. GTCS, generalized tonic–clonic seizrures; CBZ, carbamazepine; PHT, phenytoin; VPA, valproic acid; LTG, lamotrigine. ^aAdjusting for data source, gender, and age at death. ^bAdjusting for data source. ^cThis category includes PHT, LTG, CBZ, and/or VPA therapy.
GTCS frequency per year
0	1.00 (Ref)	1.00 (Ref)	1.00 (ref)	1.00 (ref)	1.00 (ref)	1.00 (ref)	1.00 (ref)
1–2	5.1 (3.0–8.5)	5.0 (3.0–8.6)	4.7 (2.7–8.0)	5.0 (2.9–8.6)	4.9 (2.9–8.4)	5.0 (2.9–8.6)	6.4 (3.4–12.0)
≥3	15.3 (10.0–23.5)	15.6 (10.1–24.2)	13.8 (8.8–21.6)	15.2 (9.8–23.9)	14.2 (9.1–22.3)	14.0 (8.9–22.0)	15.5 (9.2–26.0)
Unknown	5.1 (3.1–8.4)	5.1 (3.1–8.5)	5.2 (3.1–8.6)	5.1 (3.1–8.5)	4.9 (3.0–8.1)	5.3 (3.2–8.9)	2.3 (1.2–4.5)
CBZ therapy
No AED	1.00 (Ref)		1.00 (Ref)
Other AED	1.1 (0.6–2.0)		0.7 (0.3–1.3)
CBZ monotherapy	0.7 (0.3–1.4)		0.5 (0.2–1.1)
CBZ polytherapy	2.8 (1.5–5.3)		1.4 (0.7–2.7)
PHT therapy
No AED	1.00 (Ref)			1.00 (Ref)
Other AED	1.2 (0.7–2.1)			0.8 (0.4–1.5)
PHT monotherapy	0.7 (0.3–1.6)			0.4 (0.2–1.1)
PHT polytherapy	2.1 (1.1–4.2)			0.98 (0.5–2.1)
VPA therapy
No AED	1.00 (Ref)				1.00 (Ref)
Other AED	1.2 (0.7–2.2)				0.8 (0.4–1.5)
VPA monotherapy	0.5 (0.2–1.1)				0.4 (0.2–1.01)
VPA polytherapy	2.6 (1.4–5.0)				1.2 (0.6–2.5)
Other therapy
No AED	1.00 (Ref)					1.00 (Ref)
Other AED^c	0.9 (0.5–1.7)					0.6 (0.3–1.2)
Other AED monotherapy	1.1 (0.4–3.0)					1.1 (0.4–3.2)
Other AED polytherapy	2.8 (1.5–5.3)					1.4 (0.7–2.8)
Number of drugs
0	1.0 (Ref)						1.0 (Ref)
1	0.6 (0.3–1.2)						0.5 (0.3–0.995)
2	1.5 (0.8–2.8)						0.9 (0.4–1.8)
3	3.8 (2.0–7.4)						2.0 (0.9–4.1)
>4	4.0 (1.7–9.3)						1.6 (0.6–4.1)

To examine whether the English study was responsible for these findings, we examined the results separately in the English study and in the combined U.S. and Swedish studies. Results were unchanged in the English study. In the combined U.S. and Swedish studies, the estimates for number of AEDs did not converge, although number of GTCS remained a risk factor.

In a separate analysis of lamotrigine (Table 2), SUDEP was associated with lamotrigine polytherapy and with yearly GTCS frequency on crude analysis. The association between lamotrigine and SUDEP, however, disappeared when adjustment was made for frequency of GTCS, age at death, gender, and study. What remained was a strong association between GTCS frequency per year and SUDEP. In separate adjusted analysis of the association between lamotrigine and SUDEP stratified by gender, lamotrigine monotherapy was associated with an OR of 0.4 (95% CI 0.1–2.8) in male patients and an OR of 6.6 (95% CI 0.3–174.9) in female patients. Lamotrigine polytherapy was associated with an OR of 0.8 (95% CI 0.3–2.4) in male patients and an OR of 2.0 (95% CI 0.3–13.2) in female patients. For both lamotrigine monotherapy and lamotrigine polytherapy, CIs overlapped in male and female patients and no associations were statistically significant.

Table 2.   Odds ratios and 95% confidence intervals for lamotrigine, primary or secondary generalized tonic–clonic seizure (GTCS) frequency, and risk for SUDEP data from England and the United States

Factors	Crude analysis	Adjusted analysis
		GTCS	LTG and GTCS
N = 829.
GTCS frequency per year
0	1.00 (Ref)	1.00 (Ref)	1.00 (Ref)
1–2	6.8 (3.7–12.5)	6.9 (3.7–12.8)	6.8 (3.7–12.6)
≥3	19.1 (11.8–31.0)	19.6 (12.0–32.0)	19.3 (11.6–32.1)
Unknown	2.3 (1.2–4.4)	2.4 (1.3–4.6)	2.4 (1.2–4.5)
LTG therapy
No AED	1.00 (Ref)		1.00 (Ref)
Other AED	1.1 (0.6–1.9)		0.7 (0.4–1.4)
LTG monotherapy	1.5 (0.4–6.2)		0.7 (0.1–3.6)
LTG polytherapy	2.9 (1.4–6.0)		0.95 (0.4–2.2)

Discussion

In our previous work (Hesdorffer et al., 2011), we opined that decreasing GTCS frequency may be more important than decreasing the number of AEDs in reducing SUDEP risk. Our current analysis underscores the importance of GTCS control and suggests that the type and number of AEDs do not increase SUDEP risk when yearly GTCS frequency is also considered, thus implying that the number and type of AEDs simply reflect the severity of epilepsy. Indeed a meta-analysis of add-on drug trials found higher SUDEP rates in placebo groups who had fewer AEDs but worse seizure control (Ryvlin & Rheims, 2009).

Each of the individual case–control studies considered here had consistently shown that GTCS frequency was associated with an increased risk for SUDEP, and two have shown that the number of AEDs remained significantly associated with SUDEP after adjustment for GTCS frequency (Nilsson et al., 1999b; Walczak et al., 2001). When three or more AEDs were used, the number of GTCS and the number of seizures were simultaneously considered in the U.S. study, and all three remained associated with an increased risk for SUDEP (Walczak et al., 2001). Similar findings were reported in the Swedish study for GTCS frequency and three or more AEDs (Nilsson et al., 1999b). In contrast, we found that number of AEDs was no longer associated with SUDEP after adjustment for GTCS frequency and that neither were specific AEDs alone or in polytherapy. In our analysis, GTCS was used to denote primary or secondary generalized seizures, according to the previous classification. Separate analysis conducted in the English cohort and in the combined Swedish and U.S. cohorts did not illuminate whether the finding was due entirely to the larger English study.

The findings of our combined analysis are consistent with case reports and animal models that suggest that SUDEP is most often related to a GTCS as is further illustrated in video–electroencephalography (EEG) recordings of SUDEP and near SUDEP in epilepsy monitoring units. These cases underscore that GTCS triggers the cascade of events leading to death, and suggest three primary, possibly interrelated, seizure-induced mechanisms: respiratory failure, cardiac arrhythmia, or cerebral electrical shutdown (Tomson et al., 2008; Surges et al., 2009; Bateman et al., 2010; Lhatoo et al., 2010).

Carbamazepine and lamotrigine have been identified as SUDEP risk factors in some previous case–control studies (Tomson et al., 2008; Hesdorffer et al., 2011), and lamotrigine in idiopathic generalized epilepsy has been suggested as a SUDEP risk factor in case series (Aurlien et al., 2007). Our finding that none of the AEDs considered were associated with SUDEP risk as monotherapy or as polytherapy when GTCS frequency was taken into account provides a consistent message that number of GTCS increases SUDEP risk and not AEDs. In addition, the separate analysis of male and female patients failed to find a significant association between lamotrigine and SUDEP. Our results suggest that prevention of SUDEP must involve increased efforts, even including intensified pharmacotherapy, to decrease GTCS frequency in order to avert the occurrence of this devastating epilepsy outcome. Our results also suggest that future studies of SUDEP risk factors must be large enough to adjust for confounders and that matching on GTCS frequency is crucial for uncovering other SUDEP risk factors that may be of smaller magnitude.

Disclosures

Ettore Beghi: Pharmaceutical companies supporting studies are GlaxoSmithKline, EISAI, Sanofi-Aventis, and Sigma-Tau; Emma Benn: None; Martin Brodie: Scientific Advisory Boards of Pfizer, UCB Pharma, Eisai, GSK, Novartis, Valeant, Sierra Neuropharmaceuticals, and Medtronic. Speaker honoraria from UCB, Eisai, GSK, Novartis, and Valeant; W. Allen Hauser: Consultant for Pfizer and GlaxoSmithKline; Dale Hesdorffer: Advisory board for Pfizer; Yvonne Langam: None; Lena Nilsson: None; Josemir W Sander: Honoraria, consultancy fees, and research grants from various pharmaceutical companies including UCB, Eisai, Janssen-Cilag, and GlaxoSmithKline; Torbjorn Tomson: Research grants and/or speakers honoraria from Eisai, GlaxoSmithKline, Janssen-Cilag Novartis, Sanofi-Aventis, Pfizer, and UCB-Pharma; Thaddeus S. Walczak: None. The authors confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.