Does Second Impact Syndrome Exist?

It has often been reported that when a person suffers a second concussion (mild traumatic brain injury) before recovery from the first concussion takes place, that catastrophic brain swelling and death can result. This has been alleged to occur based on anecdotal information and when it happens (usually in young male athletes) it is referred to as “second impact syndrome” (SIS). SIS is often reported as a factual entity in the media and in the medical literature.


However, when one assess the evidence-based data for whether SIS truly exists, we are left with far more questions than answers. As Dr. Paul McCrory and colleagues (2012) put it, “The phenomenon of the second impact syndrome (SIS) continues to appear in the medical literature in spite of the lack of systemic evidence for its existence.” While McCrory's work is accessible through medical libraries, I have never seen in discussed in mainstream media coverage of SIS. Thus, a brief summary is presented below for ease of access for those who are interested.

Dr. McCrory has been studying SIS with evidence-based methods for over a decade. In fact, the seminal empirical study on this topic was performed by McCrory and Berkovic (1998). This was followed by a classic 2001 paper by McCrory entitled “Does Second Impact Syndrome Exist?,” which is the basis of the title of this blog entry.

In the 2001 paper, McCrory noted that 35 cases from the National Center for Catastrophic Sport Injury (NCCSI)  were cited as probable cases of SIS but not published in scientific journals due to lack of confirmatory details. McCrory referenced his 1998 study, in which he and Berkovic independently applied four basic criteria to 17 published SIS cases to determine how many would result in a diagnosis of definite SIS. Those criteria were as follows:

1) Medical review after a witnessed first impact.

2) Documentation of ongoing symptoms following the first impact up to the time of the second impact.

3) Witnessed second head impact with a subsequent rapid cerebral (brain) deterioration.

4) Neuropathological or neuroimaging evidence of cerebral swelling without significant intracranial hematoma (bleeding) or other cause for edema (swelling).

Not a single patient met all four criteria. Thus, there was not a single confirmed case of definite SIS in any of the cases reviewed despite the fact that all of them had been previously described or quoted as examples of SIS. Further analysis of the study revealed even more interesting details. For example, in 11 of 17 published cases, there was no evidence that a second impact actually occurred despite the fact that four of those cases had been heralded as “classic” examples of SIS in the literature. They also found that most of the evidence used to support a first impact was based on teammate recall (subjective evidence) rather than video tape (objective evidence).

Rather than merely accepting that the teammate recall was correct, McCrory and Berkovic performed a study of 102 football players to assess the reliability of their recall for concussive injuries sustained by their teammates. Their recall was compared to video-taped reviews and recall of the injured player. McCrory and Berkovic found that the teammates significantly over-reported the presence of initial impacts. In other words, they tended to report that a concussion occurred when objective evidence indicated otherwise. In only 11% of cases was there immediate medical documentation after the reported initial impact. Seventy-one percent of the cases were described as not meeting SIS criteria, whereas five of the 17 cases (29%) met three of the four criteria, with neither having a medical review after a witnessed first impact. These cases were described as probable (but not definite) SIS.

In McCrory’s 2001 article, he questioned why it is that SIS only tends to be reported in the United States despite the fact that similar or greater concussion rates occur in competitive sports throughout the world. To research this issue further, McCrory (who is from Australia) analyzed all deaths due to Australian football from 1968-1999. In these 32, there were 25 player deaths. Of these, nine were due to a neurological trauma and the others were due to non-neurological factors. Of the nine traumatic brain injury deaths, how many met criteria for SIS? None. No cases of SIS despite the fact that McCrory notes that the concussion risk in Australian football is about eight times that of American football. The risk of death from brain injury in this sample was 1 in 30 million player games. McCrory also questioned why there are not more cases of SIS in boxing compared to other sports due to the frequent concussive head trauma experienced in most fights.

McCrory also stated in his 2001 paper was that the published 1984 case by Saunders and Harbaugh that first used the term “second impact” to describe the death of a 19-year-old college football player actually did not have a described second impact. It is true that the wording used by Sanders and Harbuagh did not describe the second impact but it is also true that their wording does not preclude that some type of second impact to the head occurred. Specifically, all that they say about the second impact is as follows, “Despite accounts of no unusual head trauma, he walked from the field and collapsed.” The key word is “unusual.” This technically does not mean that no head trauma occurred but it means that there was no head trauma that anyone observed that was out of the ordinary. One could argue, however, that a head injury significant enough to cause a concussion would be considered unusual.
  
McCrory noted that belief in SIS has reached “almost mythical proportions” and that the term “second impact syndrome” is misleading. He suggested that SIS be replaced with the term “diffuse cerebral swelling” (a finding that was present in most of the cases he reviewed), and that this extremely rare condition is caused by a single brain injury (emphasis mine). I strongly agree with him. Another point I agree with him is the following conclusion from his 2001 paper:

“Most cases of traumatic cerebral swelling, whether associated with a structural brain injury or not, have no prior evidence of head injury with ongoing symptoms that would support the concept of second impact syndrome as defined in the literature. In those cases that are presumed to represent SIS, the evidence that a prior head injury is a risk factor for this pathophysiological entity is not compelling.”
 
One of the other points McCrory makes in his 2001 article relates to when the initial clinical presentation does not match the objective severity of the injury because of a gradual progressive decline as opposed to an immediate and rapid one. He noted that scientists have commented on this phenomenon dating back to 1891 and that the condition is now commonly referred to as “talk and die syndrome” because the person is initially conversant but then gradually and fatally deteriorates (or comes close to fatally deteriorating). He noted that this delayed response occurs in about 15% of all cases of “severe” head trauma and is caused by an intracranial hematoma (bleed) in 75% of those cases. Sometimes, the bleed is present when the patient’s brain is initially scanned but sometimes the bleed is not observable until after a delay.

It has been over 10 years since McCrory published his 2001 paper and so I contacted him to determine if his views on SIS had changed. McCrory’s response to me on August 11, 2010 was as follows and I am reporting it for the first time here (with permission):

“My conclusions on 'SIS' are unchanged. I have continued to track the literature and claims made on this topic and while I am happy to accept that perhaps someone may be able to satisfy the criteria (as per the original article) and really have the condition, I am yet to see it.  It is worth observing that the US is the only place on the planet where SIS is actually discussed as an entity. Most other countries have moved on from this viewpoint years ago.”

Yet, in the U.S., the most recent (June 2012) Guidelines for Concussion Management in the School Setting, published by the New York State Education Department (SED), states as a fact that "Additionally, children and adolescents are at increased risk of protracted recovery and severe, potential permanent disability (e.g. early dementia also known as chronic traumatic encephalopathy), or even death if they sustain another concussion before fully recovering from the first concussion."  It's a classic example of mythology prevailing over data. As a disclaimer, I served as a formal advisor to SED regarding this document and have asked to be removed due to statements such as these (and others that I strongly oppose) making their way into the final document despite presenting the data described above. There is not even mention of at least a controversy being present about this issue.

McCrory and colleagues (2012) and Randolph and Kirkwood (2009) point out that delayed cerebral swelling does not require multiple injuries to occur, is most likely caused by a genetic susceptibility, and is extremely rare. For example, Randolph and Kirkwood cited statistical data showing that the rarity of this outcome in 10 years of American football (1997-2006) was 1 in 1.8 million players. Despite this rarity, millions of dollars are being spent on baseline cognitive testing, largely with the hopes of preventing this outcome. People are scared by the media reports and schools do not want to be held liable. Hence, the need to want to “do something” and testing fits the bill. The problem is that the bill (using conservative estimates), according to Randolph and Kirkwood, amounts to 36 million dollars over 10 football seasons to try and prevent a 1 in 1.8 million person outcome.

McCrory and I agree that it makes sense to hold patients out of play after they have suffered a concussion (consistent with the 2009 Zurich consensus guidelines on return to play after concussion; which were authored by McCrory) because it reduces that chances of further injury to the brain and/or other parts of the body (due to slower reaction time) but not because it is going to reduce the risk of post-traumatic cerebral swelling. I also agree with Randolph and Kirkwood that the best way to manage patients who suffer head trauma during sports (and to try and reduce a catastrophic outcome) is via close observation to detect a neurological decline and to immediately initiate medical/neurosurgical intervention if such a change is detected.

Related articles:

 1. Junior Seau Did Not Suffer 1500 Concussions
 2. CTE and Suicide Link is Premature Speculation

References:

McCrory PR, Berkovic SF. Second impact syndrome. Neurology; 1998;50(3):677-683.

McCrory P. Does second impact syndrome exist? Clinical Journal of Sport Medicine; 2001;11(3):144-149.

McCrory P, Meeuwisse W, Johnston K, et al. Consensus statement on concussion in sport - the Third International Conference on Concussion in Sport held in Zurich, 2008. Phys Sportsmed. 2009;37(2):141-159.

McCrory PR, Davis, G, Makdissi, M. Second Impact Syndrome or Cerebral Swelling after Sporting Head Injury. Current Sports Medicine Reports; 2012; 11(1),21-3.

Randolph, C. & Kirkwood, M. What are the real risks of sport-related concussion, and are they modifiable? Journal of the International Neuropsychological Society; 2009;15, 512–520 .

Saunders RL, Harbaugh RE. The second impact in catastrophic contact-sports head trauma. JAMA; 1984; 252(4):538-9.