from FEAT:
Autism Tragedies, Autism Inquest
[This special report is from Jane Pickett of the Autism Tissue Program Research Park Princeton, NJ.]
Dennis Debbaudt summed up the sad news this week about the little boy of five who was missing and then found drowned in Pompano, Florida, ‘Body of missing autistic boy found in Pompano canal
http://www.sun-sentinel.com/news/local/broward/sfl-0210missingboy.story?coll=sfla%2Dnews%2Dbroward
“This all too common story illustrates yet another reason to get ourselves and others educated about the risks associated with kids who have autism”. Dennis is the author of “Avoiding Unfortunate Situations”, the parent of a young adult son with autism, and a 25-year veteran of law enforcement in the private sector who has developed police training materials and has presented this information to over 1,000 law enforcement professionals.
He is absolutely correct about the need to assess risk in the autism population. A publication last month by Sherrard, Tonge and Ozanne-Smith Injury Risk in young people with intellectual disability”, Journal of Intellectual Disability Research 46:6-16), reported on risk assessment in young people with ID (Intellectual Disability) in Australia.
Their abstract states:
Australia study demonstrates a largely unrecognized public health problem of a high unintentional injury risk for young people with ID, and identifies substantial and important risk factors. Injury prevention programmes with a particular focus on this with highly disturbed behaviors or epilepsy, are warranted in this population. Disruptive, self-absorbed, have communication disturbance, are anxious, have problems relating socially or have ADHD symptoms are at increased risk for injury. Age sex and IQ not sign risk factors – a result in sharp contrast with the general pop where injury incidence varies with age and males have approx. double the injury risk of females (Gallagher 1984).
Autism Risk and Mortality - An Inquest is Needed
I have received many questions about the recent publication, ‘Causes of Death In Autism’, written with colleagues Drs. Robert Shavelle and David Strauss. We report a higher mortality rate among those with autism. What follows is the rationale for the study was done and its conclusions.
There were several reasons that prompted this study. One was a question from Dr. Bernie Rimland, who asked if the Autism Tissue Program had autism mortality information, citing a reference for a two-year-old study by Strauss and Shavelle reporting data from the extensive California developmental disabilities registry, with 11,347 autistic individuals, that clearly showed that persons with autism are subject to increased mortality risk, with an overall mortality ratio of 213% (mortality ratio, MR, is observed deaths compared to expected deaths converted to a percentage). The MR for females, 490%, was notably higher than that for males, 167%. The life expectancy of a 5-year-old in that cohort of was reduced by 6.1 years for boys and 12.3 years for girls.
About the same time, in the summer of 2000, the Autism Tissue Program was contacted for the brain donation for a nine-year-old boy, who died suddenly and unexpectedly. Earlier that year, a girl, 7, and a boy, 8, died in other states. In September, another boy, 8, died at home in Minnesota. All had a seizure disorder; however, the disorder was not considered life-threatening, and, the very first donor to the Autism Tissue Program was a child of 14, with no known seizure disorder, died suddenly and unexpectedly of hyperthermia. These deaths raised questions about mortality. Also, in the summer of 2000, I met with local autism advocacy leaders at the M.I.N.D. Institute to discuss mutual brain research goals with the Autism Tissue Program. In support of this collaborative relationship between NAAR, the ASA Foundation and the M.I.N.D. Institute, the East Bay chapter of ASA gave the Autism Tissue Program a donation of$10,000. These funds were spent to further analyze California mortalitydata. The conclusions of the study follow (references and/or reprints supplied on request – email [email protected]):
‘The aim of this study was to expand upon the initial report of excess mortality in the California autism population (Strauss & Shavelle, 1998) by exploring the cause of the death. We excluded the non-ambulatory, a group known to have a very high mortality rate; even so, subjects had higher mortality than expected in the general population, after adjustment for age and sex.
As medical records on cause of death were not generally available, we worked with computer codes and death certificates. It should be noted that the cause? Specific SMRs reported here are conservative, as in 16% of the cases we were unable to determine the cause. Although these sources are frequently used in large epidemiologic studies, their limitations are well known (Lilienfeld & Lilienfeld, 1980). Resulting bias may have been reduced, however, because the same source (California mortality computer tapes) was used for the causes of death of both the subjects with autism and the reference California population.
In the no or mild mental retardation group, deaths by seizures, nervous system dysfunction, drowning, and suffocation were all more than three times higher than would be expected in the general population. For the more severely retarded subjects, all categories of cause of death except cancer and “all other causes >800” were more than three times higher. External causes of death excluding drowning and suffocation, which were analyzed separately, were actually lower than expected. This is largely because the general population death rate from injuries climbs rapidly in the teen years, whereas many autistic teenagers may lack the same opportunities for risky behavior.
Overall, children between the ages of 5 and 10 had the highest SMRs, especially girls with an SMR of 16.8. Further research is needed on why some causes are elevated, and on the male/female/age differences. Future studies might also examine medical records and death certificates. The present study reflects a fifteen-year period ending in 1997. Given the large increase in autism intakes into the California system from the first to the second half of the last decade (148%, unpublished data), a follow-up through 2000 should provide further information on the younger age groups.
State databases and individual records are extremely important in obtaining the numbers necessary to examine cause of death, or other characteristics of interest in autism. For instance, the prevalence of epilepsy for all ages in this study was 4.2 %. This figure is at the lower end of the range of seizure frequency (4-60%) reported in various studies. The CDER contains fields for 8 seizure types, frequencies for each, condition impact and whether the person takes anticonvulsant medication. Further data mining could assess age, level of cognitive function and type and degree of language dysfunction pertinent to epilepsy (Tuchman, 2000).The potential of these data sources is immense and work should continue to improve the accuracy and consistency of the autism data in all states.
The prerequisite for accuracy of data is equally true for mortality reporting. In the present study, a review of ICD-9 codes for the 202 deaths showed many instances of ill-defined causes of death. For example, 3 were coded 299 (infantile autism); 13 coded 318 or 319 (moderate mental retardation or mental retardation not otherwise specified); and 16 were coded 799 (unspecified causes). It is estimated that sudden unexplained deaths (SUD) claims over 4000 persons between the ages of 1 and 22 each year in the U.S. (Ackerman et.al., 2001). Ackerman and colleagues (1999, 2001)used post-mortem ‘molecular’ tissue autopsies to link a KVLQT1 gene defect with fatal cardiac arrhythmias associated with one form of inherited long QT syndrome (LQTS) and swimming-triggered cardiac events. Another autosomal-dominant LQT gene, HERG, is located on chromosome 7q35-36 (Wanget. al., 1998).
Deaths due to drowning and epilepsy are common in the California autism population and deserve further forensic study as does the possibility of ‘acquired’ LQTS from various medications. Ponti et. al (2001) have called for a consensus to organize information on QT prolongation and occurrence of Torsades de Pointes, polymorphic ventricular tachycardia thought to be initiated by early after-depolarizations in the cardiac Purkinje system (Chiang & Roden, 2000), with non-anti arrhythmic drugs.
Looking strictly at patient populations with epilepsy, the rate of sudden, unexpected and unexplained death in epilepsy (SUDEP) exceeds the expected rate of sudden death in the general population by nearly 24 times (Ficker, et. al. 1998; Annegers & Coan, 1999). The incidence and risk of SUD and SUDEP will only be fully ascertained with careful postmortem examination and complete death narratives.
The information from this study is relevant to the Autism Tissue Program (ATP) that recruits brain donors for biomedical research, as it reveals age-trends in the number and causes of death. Without this data, one would not know if a given group of brain tissue cases was representative of deaths in the autism population, or simply an unusual sample. For example, the twenty-three donations through June 2001, 43% of which were under the age of 15, may now be viewed as fairly typical. The causes of death were: seizures (6, 2 of these in the SUDEP category), drowning (5), external injury such as auto trauma (4), congestive heart failure (2), hyperthermia (1), infection (1), GI bleeding and seizure disorder (1), anoxia (1), sepsis from infected bowel (1), cardiac arrest, SUD (1). Nevertheless, each donor represents one or more potential subsets of autism characteristics orco-morbidity that must be taken into account in brain tissue research. Hyperthermia was associated with at least three deaths and, in two of these, neuroleptic malignant syndrome (Andreassen & Pedersen, 2000; Kahn & Farver, 2000) was thought to contribute to the cause of death.
While the annual monetary cost of autism in the US is estimated at roughly $26 billion (Maltby, 2000), the impact of higher mortality is an incalculable emotional cost to families. There should be a concerted effort to address autism aging, morbidity and mortality. We hope that this initial study on causes of death in autism will lead to improvements in prevention, care, and planning for this population.’ (end of conclusions to report)
Risk and Mortality Assessment – An Inquest is Needed
Like many studies, more questions have been generated than were answered. It should; however, help promote an inquest into autism injury risk and deaths in California, the US and even other countries. It is important to define the magnitude of the problem and identify the associated causes and risk factors that contribute to mortality. There is an excellent lecture on the subject of risk assessment on the web(http://www.pitt.edu/~epi2670/) by University of Pittsburgh Epidemiology Professor, Thomas J. Songer, Ph.D. This discussion of monitoring and risk assessment comes from that site. Death and injury monitoring, or surveillance, is very important as a means to formulate reasonable injury and death PREVENTION initiatives. Other injury control measures such as airbags, helmets, seatbelts were implemented only after it became apparent that injuries were occurring in relatively high numbers. Effective injury control is based upon an understanding of injury rates; i.e. how many injuries occur within a given population. The primary means to identify injuries is to establish a surveillance system that is generally defined as the systematic assessment of disease or injuries with a common method. Monitoring occurs over time and thus allows researchers to evaluate changes. Death certificates, medical records, surveys of medical facilities, coroner and police reports all contain different and potentially useful information.
The first data source often examined in injury research is death certificates. In the United States, the National Center for Health Statistics (NCHS) maintains information on deaths and death certificates. The NCHS gathers this information from the individual states. Each state, in turn, has a law that requires for death certificates to be completed for all deaths. A standard death certificate form is collected across the states. Thus, the death certificate data compiled by the NCHS represents one of the few sources that allow for comparisons across geographic regions and standard demographic factors. Unfortunately, it is difficult to obtain adequate cause of injury or death information. Moreover, according to Dr. Songer, existing data systems are often difficult to integrate together. Crime-based systems do not always assess injuries and emergency medical transport systems do not always match up with emergency department or hospital admission records.
Active studies employ regular outreach and monitoring of surveillance sites. And what are these surveillance sites? They are most often state and federal governmental surveys. One in particular that could provide information is the CDC Behavioral Risk Factor Surveillance System that keeps statistics on asthma, diabetes, alcohol consumption, etc., but not autism. Autism can be monitored if it becomes a ‘reportable’ condition - the CDC reporting system is vast and might already have the capacity to monitor autism injury and deaths.
The NCHS is also responsible for the government-based health and health care surveys in the United States. Two important surveys for injuries are the NHIS (National Health Interview Survey) and the NHAMCS (National Hospital Ambulatory Medical Care Survey). The NHIS is used primarily to monitor the health of the population. It includes several questions on physician diagnosed illnesses and disability. It is based upon a sample of the population and represents a cross-sectional survey. It is conducted annually. Injury-specific items are included in the NHIS. Two questions identify injury and poisoning events. Both address events which occurred in the previous 3 months and which required medical attention. The recent redesign of the survey has added further information on injuries. Data are now collected on external causes of injuries (E-codes), and the circumstances that surround the injury (type, activity, place of occurrence).
A question is how useful is this data in 2002? The data in the NHIS apparently represent only self-reported injuries. National Health Interview Survey on Disability (NHIS-D) was last updated on 9/16/98. The NHAMCS is a survey from the NCHS. It assesses the characteristics of ambulatory care use in hospital settings. This included emergency department activities. Thus, it is of interest to injury researchers. The NHAMCS is not an annual survey, though, and the latest data available are for the years 1995-96.
Somehow, we need to begin the process of standardizing the information that is collected on injuries and deaths in children with autism ad related disorders. In this way, it will be possible to have some comparisons between data systems and develop responsible prevention programs.
Regarding brain tissue donation; no, it is not counter-productive to the Autism Tissue Program to encourage prevention. It is estimated that around 4-5000 persons with autism die annually (6300 Americans die each day,2.3 mill each year US). If even 1% of the families chose to donate brain tissue after their loved one died, there would be almost more cases in a year than exist now. Every unexpected death of a child will prompt an investigation and autopsy; the brain and other organs will be removed and examined. If not donated, the brain is returned to the body cavity and funeral arrangements take place (viewing is possible). Also, being an organ donor does not mean that brain tissue is automatically donated. It is necessary to call 1-877-333-0999 for brain donation.
Everyone I work with at the Autism Tissue Program values quality of life, protection of children and adults when needed, and research to understand brain function and autism.
Most sincerely,
Jane Pickett, Ph.D. Director,
Autism Tissue Program 99 Wall Street,
Research Park Princeton, NJ 08540
1-877-333-0999 (24-hour) Fax: (609)430-9163
Email: [email protected]
Web: www.brainbank.org