Postmortem Genetic Testing Is an Increasingly Utilized Tool in Death Investigation

Abstract

Introduction:

Postmortem genetic testing (PMGT) can provide valuable information about an individual’s cause of death and potentially allow at-risk relatives to discern their risks for inherited cardiac disease. Postmortem genetic testing is most often successful with certain specimens.

Methods:

Investigators collected data on postmortem referrals to GeneDx, LLC for PMGT. Orders were reviewed and stratified based on provider, specimen type, and tests ordered.

Discussion:

This cohort included 601 deceased individuals referred for PMGT with a total of 673 genetic tests ordered from 247 different providers. The most common test categories ordered were arrhythmia (33.4%) and cardiomyopathy (29.3%). A likely pathogenic or pathogenic genetic variant was identified in approximately 15% of patients. Blood in EDTA was received for 21.6% of patients with a 95% success rate for completion of all test components. Blood samples in EDTA were most successful in completing PMGT, but sequencing was still successful in the majority of suboptimal specimens.

Conclusion:

The use of PMGT is increasing. Obtaining optimal samples (blood in EDTA) is important for successful completion of genetic testing. Obstacles may still exist for obtaining and storing ideal specimens. Continued efforts are needed for education and awareness around appropriate specimen types, storage and shipping of specimens, DNA banking, and overall availability of PMGT. In addition, access to resources such as supplies, proper storage conditions, DNA banking, and PMGT will allow for more opportunities to complete testing.

Introduction

Sudden death at any age is a tragic event. Occurrence in a young person can be especially traumatizing for a family and community. Sudden cardiac death of the young (SCDY) is defined as sudden death between the ages of 1 to 40 years and is cardiac in nature (1). Cases of noncoronary artery disease related sudden cardiac death (SCD) often have an underlying genetic etiology (2).

Cardiomyopathies and arrhythmias are common causes of genetic SCD. Many cardiac conditions are inherited in an autosomal dominant manner, and there is a one in two (50%) chance that first degree relatives will inherit the genetic predisposition. Testing of an affected person is the best way to identify a genetic predisposition in a family but can be difficult to achieve in sudden death cases.

In cases of SCD, postmortem genetic testing (PMGT) has the potential to identify the etiology of death in cases with cardiomyopathy, aortopathy, and negative autopsies (may be suspicious for a cardiac channelopathy). A genetic predisposition is identified in such cases about 13% to 41% of the time (3 -9). Because sudden death can be the first and only symptom of disease and other family members may be unaffected or asymptomatic, it is critical to save a sample of the decedent for genetic testing to allow for a more specific diagnosis (7).

Discovery of a genetic predisposition can provide insight about an individual’s cause of death and allows for identifying and preventing SCD in at-risk relatives (3). Postmortem genetic testing to rule-in or rule-out an inherited predisposition is cost and time effective in the postmortem setting in conjunction with genetic counseling (10). On the other hand, serial clinical cardiac screening (e.g., electrocardiogram and/or echocardiogram) without results from genetic testing on the deceased individual requires repeated follow-up of at-risk relatives for whom the disease may or may not develop (11). Furthermore, clinical cardiac screening leads to a cardiac diagnosis approximately 10% to 30% of the time, while the remainder are left without a diagnosis (12 -16).

Postmortem genetic testing including multigene panel testing, exome sequencing, or genome sequencing can be used to identify variants in genes associated with SCD. These variants are evaluated and interpreted based on available evidence from various sources (i.e., prior publications, mutation databases, functional evidence, segregation data, or frequency of the variant population controls) and classified according to the current guidelines for variant classification (17). Those variants classified as pathogenic or likely pathogenic are reported and may be considered diagnostic if they correlate with the available clinical data. The importance of such variants are also recognized by the inclusion as part of the genes recommended as reportable secondary variants for exome sequencing results (18). The vast majority of pathogenic variants associated with SCD are sequence changes in cardiac genes, such as single nucleotide variants, small insertions, deletions, or splice site changes. Nevertheless, intragenic or whole gene deletions or duplications known as copy number variants (CNVs) may also occur in any gene. For example, in long QT syndrome, CNV represent up to 10% of pathogenic variants in patients without an identifiable variant by sequencing (19, 20). Therefore, it is important to integrate CNV analysis into PMGT approaches, by performing concurrent CNV analysis from next-generation sequencing (NGS) data or by exon-level microarray analysis.

Once a pathogenic or likely pathogenic variant is identified in a postmortem case, cardiovascular clinical follow-up including genetic counseling is indicated for family members to put the results in context with other findings from the autopsy, the decedent’s personal health history, and the decedent’s family history (21). If it is deemed that the pathogenic or likely pathogenic variant could have contributed to the cause of death of the proband, cascade testing of family members is indicated. Postmortem genetic testing represents a more timely and precise approach than clinical screening to identify family members who are either presymptomatic or not at increased risk for developing the cardiovascular disorder. For example, individuals with pathogenic variants in the LMNA gene are at increased risk for ventricular tachycardia and ventricular fibrillation prior to developing signs of cardiomyopathy. Because of this finding, guidelines recommend patients with nonischemic cardiomyopathy due to LMNA pathogenic variants with two or more risk factors (non-sustained ventricular tachycardia [NSVT], left ventricular ejection fraction [LVEF] <45%, non-missense pathogenic variant, and male sex) be considered for an implantable cardioverter-defibrillator (ICD) if survival of greater than one year is expected (22). Identification of a pathogenic LMNA gene variant in at-risk family members is essential to clarify their risk for arrhythmia and/or cardiomyopathy and can significantly alter medical management in these individuals. This same concept can be applied when pathogenic or likely pathogenic variants are identified in other cardiac genes associated with an increased risk for SCD.

Unfortunately, it is not always possible or easy to obtain an appropriate specimen for genetic testing from the decedent. Collection of an appropriate specimen for testing has been a long-recognized barrier to PMGT. This could be due to the fact that many institutions do not have a standard protocols, samples from certain tissues may be inadequate, and/or there is inconsistency in collecting optimal specimens. To address some of these issues, the National Association of Medical Examiners (NAME) and a group of experienced cardiovascular genetic counselors published a position paper in 2013 advocating for retaining specimens for genetic testing (7). The paper details specimen types and storage conditions needed for quality DNA extraction and successful genetic testing. Specifically, the paper states that 5 to 10 mL of blood preserved in K₂EDTA (usually a purple top tube) should be saved in a refrigerator at 4 °C for storage up to four weeks. If storage is needed for longer than four weeks, the specimen should be moved to a −20 °C to −70 °C freezer and remain frozen until the specimen is requested. Frozen specimens should be shipped on dry ice to a genetic testing laboratory. If specimens are not desired for PMGT, then then should be properly discarded according to the office’s sample retention schedule.

Since publication of the paper, the use of PMGT as a tool for death investigation has increased. Genetic counselors specializing in cardiology are uniquely adept at working with families, medical examiners, and coroners to obtain appropriate samples and facilitate PMGT (23). Many medical examiners have embraced this position paper and are more routinely saving appropriate samples for PMGT in SCDY cases. If initial attempts at PMGT do not identify a genetic cause of death, specimen banking is an option that makes repeat PMGT possible. Repeating PMGT at a later date when more is known about genotype–phenotype correlations may provide previously unattainable answers. Furthermore, if initial attempts at genetic testing are unsuccessful for technological reasons, future methodologies may be successful on a banked specimen.

The experience of a commercial laboratory, GeneDx, LLC, offering diagnostic genetic testing for cardiovascular genetic conditions including sudden death is described here. This study looks at a large cohort of referrals for postmortem cardiovascular genetic testing. This cohort is used to illuminate the current landscape PMGT from the unique perspective of a genetic testing laboratory.

Materials and Methods

Investigators collected data on postmortem referrals to GeneDx, LLC, for cardiovascular genetic testing. This study was conducted under an IRB protocol approved by the Western Institutional Review Board, Study Number 1169768, WIRB Pro Number 20162523, which states that this research meets the requirements for a waiver of consent.

Patients described as deceased by the ordering facility and for which the laboratory received at least one cardiovascular test order were included in this study. Data extracted from case requisitions included de-identified demographic information on patients and provider (e.g., patient date of death, referring facility). Order data was extracted from the laboratory database (e.g., specimen type, genetic test ordered).

Each patient and provider were labeled with a unique ID to determine the number of referred patients and ordering providers. The number of tests ordered was determined based on the exclusive test name assigned by the laboratory. Of note, a single provider referred about half of the patients in this cohort. Analyses were completed with and without referrals from this large single referring provider. With the exception of specimen type distribution, results did not significantly change with the inclusion of these referrals. Therefore, they were not excluded from overall analyses.

Repeat specimens received and accessioned as a separate test were excluded from most analyses because the patient and associated test(s) were already included in the dataset with the initial order. The laboratory asked for a repeat specimen when results could not be obtained because of poor quality and/or quantity of DNA from the initial specimen(s). Repeat specimen data was only used to determine the success rate of repeat specimens to deliver results.

Some referrals included more than one specimen type. Specimen(s) from a shipment were categorized and prioritized in the following order: blood in ethylenediamine tetraacetic acid (EDTA) only; blood in EDTA plus an additional specimen type; blood in other collection tube; blood in other collection tube plus another specimen type; dried blood spot (DBS) only; DBS plus other blood or specimen type; pre-extracted DNA; tissue; other. Other blood specimens (non-EDTA) were provided in red, gray, blue, green, and white top collection tubes. The tissue types received and tested included brain, heart, kidney, liver, lung, lymph node, skeletal muscle, skin, and spleen. Priority order was determined by the preference of the testing laboratory, grounded in their experience with successful DNA extraction from nontraditional specimens. The cardiovascular test menu available to providers during the time of this study is provided in Table 1 . The test menu, including the number of genes included in each test, offered by the commercial laboratory changed over the course of this study due to the ever-evolving knowledge about genotype–phenotype correlations. Test menu updates are a standard laboratory practice.

Table 1:

Profile of Cardiovascular Genetic Tests by Phenotype Offered by a Commercial Laboratory Used for Postmortem Testing in This Study.

Test category	Test name	Number of genes testeda
Arrhythmia	Arrhythmia Panel	Up to 46 genes
	Long QT Syndrome (LQTS) Panel	Up to 17 genes
	Brugada Syndrome Panel	Up to 7 genes
	Catecholaminergic Polymorphic Ventricular Tachycardia (CPVT) Panel	Up to 2 genes
	Short QT Syndrome (SQTS) Panel	Up to 3 genes
Cardiomyopathy	Cardiomyopathy Panel	Up to 91 genes
	Hypertrophic Cardiomyopathy (HCM) Panel	Up to 25 genes
	Dilated Cardiomyopathy (DCM) Panel	Up to 38 genes
	Arrhythmogenic Right Ventricular Cardiomyopathy (ARVC) Panel	Up to 13genes
Arrhythmia and cardiomyopathy	Combined Cardiology Panel	Up to 120 genes
Sudden cardiac arrest	Sudden Cardiac Arrest (SCA) Panel	Up to 13 genes
Marfan/Thoracic aortic aneurysm and dissection (TAAD)	Marfan/TAAD Panel	Up to 23 genes
Otherb	Hereditary Hemorrhagic Telangiectasia (HHT) Panel	Up to 4 genes
	Heterotaxy Panel	Up to 10 genes
	Single Cardiac Gene Analysis	1 gene
	Familial Cardiac Variant Analysis	1 gene

^a Indicates the maximum number of genes that may have been evaluated for a postmortem case. Exact number depends on the time when testing was performed. The gene names for the maximum number of genes offered on each test are available in the Supplemental material (Table S1).

^b Cardiovascular testing (as listed in Table 1 ) must have been ordered on a postmortem specimen to be included in this cohort. However, multiple tests were ordered in some cases. Tests ordered in conjunction with a cardiovascular test that were not specific to the cardiovascular test menu included the following: Epilepsy panel, Noonan panel, whole genome copy number variant (CNV) array, whole exome sequencing.

Testing included NGS with or without CNV analyses via microarray for all tests listed in Table 1 . Full completion of a test was defined as all components of testing meeting quality standards for reporting. Partial completion of a test was defined as only sequencing data meeting quality standards. In some cases, CNV data was not reportable due to poor data quality. Diagnostic rate was defined as number of individuals with variant(s) classified as pathogenic or likely pathogenic variant(s) at the time of the analysis over all individuals tested.

Trends in ordering for postmortem cases were compared to all referrals sent to the same commercial laboratory to determine if changes over time were unique to this cohort. The quantity of tests ordered was plotted over time for these groups, and an exponential trend line was applied to determine the coefficient of determination (R²). The coefficient of variation was also calculated for these datasets.

Results

This cohort included 601 deceased individuals referred for PMGT. The majority (n = 536) had one test ordered, while 62 had two tests ordered, and three had more than two tests ordered. The total number of unique PMGTs ordered during this time was 673. Testing was repeated for 38 cases for which a new specimen(s) was received.

Postmortem genetic testing referrals came from 247 different providers from 213 facilities. These facilities were located across 37 states as well as internationally. More than 75% of providers referred only a single PMGT case.

The average age at death for a PMGT referral was 25 ± 19 years. The most common test category ordered was arrhythmia (n = 225, 33.4%), followed by cardiomyopathy (n = 197, 29.3%; Table 2 ). Figure 1 shows the growth in the number of PMGT referrals over time. The largest increase in growth was between 2012 and 2013, with the number of referrals in 2013 growing to more than five times the amount from 2012. The coefficient of determination (R²) when applying an exponential trendline was 0.90. The coefficient of variation was 94.7%. The best fitting trendline for the growth of all referrals sent to the same commercial laboratory over an identical timeframe (unpublished data that included both deceased and living probands) was logarithmic (R² = 0.97). The coefficient of determination was 0.85 when an exponential trendline was applied for all referrals. Pathogenic and likely pathogenic variants were identified in approximately 15% of probands. Overall, the diagnostic rate varied by phenotype, with 19.3% for cases referred strictly for a cardiomyopathy-related test, followed by 18.1% for the Sudden Cardiac Arrest (SCA) Panel, and 8.0% for arrhythmia-related tests.

Table 2:

Distribution of Tests (n = 673) Ordered for a Large Cohort of Postmortem Patients.

Test category	Test name	Number of tests ordered for this cohort
Arrhythmia	Arrhythmia Panel	199
	Long QT Syndrome (LQTS) Panel	19
	Brugada Syndrome Panel	4
	Catecholaminergic Polymorphic Ventricular Tachycardia (CPVT) Panel	2
	Short QT Syndrome (SQTS) Panel	1
	Total number of Arrhythmia-related tests ordered (% of total tests)	225 (33.4%)
Cardiomyopathy	Cardiomyopathy Panel	132
	Hypertrophic Cardiomyopathy (HCM) Panel	31
	Arrhythmogenic Right Ventricular Cardiomyopathy (ARVC) Panel	25
	Dilated Cardiomyopathy (DCM) Panel	9
	Total number of Cardiomyopathy-related tests ordered (% of total tests)	197 (29.3%)
Arrhythmia and cardiomyopathy	Combined Cardiology Panela	31
	Total number of Combined Cardiology Panels ordered (% of total tests)	31 (4.6%)
Sudden cardiac arrest	Sudden Cardiac Arrest (SCA) Panel	126
	Total number of Sudden Cardiac Arrest Panels ordered (% of total tests)	126 (18.7%)
Marfan/Thoracic aortic aneurysm and dissection (TAAD)	Marfan/TAAD Panel	14
	Total number of Marfan/TAAD Panels ordered (% of total tests)	14 (2.1%)
Other	Familial Cardiac Variant Analysis	29
	Heterotaxy Panel	8
	Single Cardiac Gene Analysis	5
	Hereditary Hemorrhagic Telangiectasia (HHT) Panel	1
	Non-Cardiovascular Testsb	37
	Total number of Other tests ordered (% of total tests)	80 (11.9%)

^a This test included genes associated with both arrhythmia and cardiomyopathy; it was only available for ∼30% of the timeframe for which data was collected.

^b Tests ordered in conjunction with a cardiovascular test (n = 37) that were not specific to the cardiovascular test menu included the following: Epilepsy panel, Noonan panel, whole genome copy number variant (CNV) array, whole exome sequencing.

Figure 1:

Number of postmortem referrals to a commercial laboratory over time.

Blood in EDTA was received for 21.6% (n = 130) of postmortem cases. The distribution of specimens received—prioritization as defined in the methods—is described in Figure 2 . The proportion of blood samples in EDTA did not significantly change over time relative to other specimen types received. Regardless of specimen type, the success rate of completing (at minimum) the sequencing portion of a test ordered was 94.9%. For 6.4% of these tests, only quality sequencing data was obtained (copy number variant analysis failed). Blood in EDTA succeeded in test completion 95% of the time, followed by pre-extracted DNA (91%), blood in other collection tubes (86%), tissue (88%), and DBS (62%). In 13 (68.4%) of 19 cases where no testing could be completed on the first specimen(s) and a repeat specimen was provided, the repeat specimen allowed for completion of either sequencing or both sequencing and CNV analysis. In 6 (42.9%) of 14 cases where the first specimen(s) allowed for completion of the sequencing but not the CNV analysis and a repeat specimen was provided, the repeat specimen allowed for completion of the CNV analysis. In most cases (30/36; 83.3%) where no testing could be completed, a repeat specimen was not provided. Specimen types for these failed cases varied. Nearly half (17/36) were pre-extracted DNA, 13 were blood specimens (seven in gray-top tubes, four in purple-top [EDTA] tubes, one blue-top tube, one red-top tube), and four were tissue.

Figure 2:

A, Distribution of specimens for all postmortem cases (n = 601). B, Distribution of specimens for all postmortem cases (n = 312) when cases referred from a single, high-volume provider consisting primarily of pre-extracted DNA were removed. *Other specimens include product of conception cultured amniocytes, cultured cells, buccal/hair, and unspecified. EDTA indicates ethylenediamine tetraacetic acid (purple-top tube); DBS, dried blood spot.

Discussion

The cardiovascular genetic test selected by an ordering provider is usually indicative of the suspected cause of death. Selection of an arrhythmia panel suggests there was suspicion of arrhythmia prior to death and/or the autopsy was negative. Selection of a cardiomyopathy panel suggests there was suspicion of cardiomyopathy prior to death and/or the autopsy showed changes in the heart muscle, which could be indicative of cardiomyopathy. The arrhythmia panel and the cardiomyopathy panel were the two most frequently ordered tests (33.4%, n = 225 and 29.3%, n = 197, respectively) in this study. This finding implies (and is consistent with the well-known fact that) these phenotypes are most commonly suspected following SCD. GeneDx, LLC, also offers a Sudden Cardiac Arrest panel and Combined Cardiology panel, which were collectively requested in 156 cases (23.4%). Though patterns in ordering slightly changed over time as the laboratory expanded its test offering to include these tests, the data still clearly indicate that arrhythmia and cardiomyopathy are the leading suspected phenotypes in sudden death cases.

Blood in EDTA is the preferred specimen for obtaining quality DNA for molecular evaluation. This specimen type represented ∼20% of the overall study population and ∼40% when removing a single provider who sent mostly pre-extracted DNA specimens and accounted for approximately half of the overall study population. Regardless, based on the types of specimens submitted to the testing laboratory, the data indicates obstacles still exist to obtaining the preferred sample type in postmortem cases. It is worth noting that the proportion of preferred specimens (blood in EDTA) did not improve over the years despite heightened awareness for PMGT, indicating that awareness alone has not significantly changed the existing sample collection procedures, and more effort is needed to address this issue. Other obstacles to updating sample collections processes may be due to lack of resources including supplies, freezers, and funding.

This study confirmed that the top-performing specimen type is blood in EDTA (succeeding 95% of the time), but also found that successful sequencing is feasible from a variety of other specimens. Quality sequencing data was obtained for nearly 95% of tests ordered, and specimens included pre-extracted DNA, DBSs, blood in non-EDTA tubes (gray, red, blue, green), and tissue. There was only a small portion (6.4%) of tests ordered where sequencing was successful and copy number variant analysis failed. This data suggests that even when ideal specimens for molecular evaluation are not available, it is advantageous to contact the laboratory to discuss making attempts on less preferred specimen types. Testing of less preferred specimens can be fruitful, but it is important to note that success is not guaranteed. Complete test failure happened most often for pre-extracted DNA and blood in gray-top tubes. Nevertheless, failure also happened (though rarer) with blood in EDTA.

These findings substantiate that besides specimen type, other important factors can influence PMGT success. A unique concern is body decomposition before a specimen can be collected for PMGT, which can lead to cell lysis and degradation of DNA. Therefore, timing of collection is a critical factor for sample success. Suboptimal specimen storage and/or shipment conditions are other factors. The increased failure rate of pre-extracted DNA specimens in our cohort likely results from decomposition. Insufficient DNA may also result from non-ideal tissue sources with a low cell content, or when DNA extraction methods are used that are not optimized for the genetic testing performed. The commercial laboratory used in this study has optimized their DNA extraction protocol for their high-throughput genetic testing. Thus, such optimal conditions cannot be guaranteed for pre-extracted DNA specimens.

Although barriers seem to still exist for obtaining the ideal specimen (blood in EDTA) for PMGT, data from this study suggest that overall awareness is increasing. Following the publication of the NAME position paper (June 2013), there was a large influx of PMGT orders to the testing laboratory. By the end of 2013, the number of PMGT referrals had increased to more than five times the amount ordered in the previous year. Additionally, PMGT referrals increased at a nearly exponential rate (coefficient of determination = 0.90; coefficient of variation 94.7%) over the course of this study. This growth rate was unique to postmortem referrals as the overall rate of test referrals to the same laboratory most closely followed a logarithmic trend.

Because variants detected by sequencing are the most common causes of cardiovascular disease, this method was prioritized over copy number analysis for postmortem specimens. Therefore, if quantity and/or quality of DNA was an issue, it was often the copy number variant analysis portion of a test that could not be completed. With increasing utilization of exome and especially whole genome sequencing, which allows for simultaneous copy number detection and requires low amounts of DNA, the quantity of DNA might be less of a concern for PMGT in the future. As technology improves and new genes are discovered, it is predicted that the types of samples able to be sequenced will increase, the cost of testing will decrease, and there will be more answers for surviving family members uncovered through genetic testing.

Conclusion

Postmortem genetic testing in cases of SCD is important for determining the cause of death and also can provide results to personalize medical management and disease prevention for at-risk relatives. Blood samples in EDTA tubes were the most likely to succeed in test completion. Submission of other types of sample types were less likely to succeed and may imply that there are still obstacles to collection of optimal specimens. This study found that PMGT will identify a pathogenic or likely pathogenic variant in approximately 15% of cases. Furthermore, PMGT significantly increased over time. Since 2013, this laboratory saw a near exponential trend in PMGT orders. Nevertheless, continued efforts are needed for education and awareness around appropriate specimen types, storage and shipping of specimens, DNA banking, and overall availability of PMGT. In addition, access to resources such as supplies, proper storage conditions, DNA banking, and PMGT will allow for more opportunities to complete testing.

AUTHORS

Rebecca Latimer MMSc, CGC, GeneDx, LLC

Roles: A,B,C,D,E,1,4,6

Heather MacLeod MS, CGC, Data Coordinating Center for the SDY Case Registry

Roles: A,B,C,D,E,2,6

Lisa Castillo MS, CGC, Northwestern University Feinberg School of Medicine

Roles: A,B,C,D,E,2,6

Daniela Macaya MQC, FACMG, GeneDx, LLC

Roles: B,C,D,E,4,6

Tara R. Hart MS, CGC, GeneDx, LLC

Roles: B,C,D,E,4,6