• Chris Brook

Differentiating differential diagnoses: SIDS vs SBS (aka AHT)

Sudden unexpected infant death (SUID) is a term used to describe the sudden and unexpected death of a baby less than 1 year old in which the cause was not obvious before investigation. If the death remains unexplained after all known causes have been explored, and after a thorough autopsy and death scene investigation, then the case is diagnosed as Sudden Infant Death Syndrome, SIDS. SIDS is basically the label for “we don’t know” in these infant deaths.


The process of exploring different possible causes of illness or death is called a differential diagnosis. As different illnesses can result in similar symptoms, doctors make a range of tests to verify or rule out known potential causes. For SUID, the differential diagnosis of known potential causes includes heart problems, genetic disorders, infections, accidental suffocation, and deliberate suffocation (child abuse). Natural disease process can never be ruled out, as unknown causes cannot be identified.


Medically, it is not possible to distinguish SIDS from suffocation. Thus, if a known medical cause such as infection is not found, investigators will look for evidence of suffocation. For example, investigators will ask where the infant slept, and whether there was soft bedding or soft objects which may have caused suffocation. If sufficient evidence of suffocation is not found, the case will be labelled as SIDS.


What about deliberate suffocation (child abuse), for example by using a pillow? Again, like unknown or untested medical causes, abuse cannot medically be ruled out as a cause. There is no “test” for suffocation and the process of ruling out other possible causes is not thorough enough to definitively rule them out, and also leaves the possibility of unknown causes. Therefore, investigators need some independent evidence that deliberate suffocation was the cause before making accusations of abuse based on such ambiguous medical evidence.


So in cases of sudden and unexpected infant death (SUID) where no known medical cause can be identified, and there is no explicit evidence of suffocation or child abuse, the cause is determined to be “unknown” and labeled as SIDS. A lack of knowledge is acknowledged. Significant resources are then invested in research to try to find the causes of these infant deaths, and to prevent them.


Some infants also die each year and present with retinal haemorrhage, subdural haemorrhage, and encephalopathy (retino-dural haemorrhages in infants, or ReDHI). This “triad” of symptoms can develop in the context of many conditions. The differential diagnosis includes medical conditions such as meningitis, sepsis, stroke, cardiac arrest, birthing trauma, coagulation and haematological disorders, metabolic disorders, hydrocephalus and congenital cranial malformations.


ReDHI (or the “triad”) can also develop in response to a traumatic event. Sometimes trauma is clearly the cause, such as the infant being in a car accident or fall. These trauma cases generally have accompanying bruising, external bleeding, and/or broken bones, as well as witnesses. Abusive trauma can also lead to such injuries, including punches or throwing an infant to the ground. Such abusive trauma also usually leaves external bruising or broken bones, and usually also has corroborating evidence. Police can be called to investigate such cases and may find evidence of the abuse, such as a witness.


In other cases, infants present with ReDHI (or the “triad”), but have no outward sign of trauma: no bruising, no bleeding, no broken bones. In 1971, Norman Guthkelch theorized that “whiplash injuries”, or shaking, may cause such injuries. However, this was just a proposal and there was no evidence to support this idea. In the subsequent 50 years there has been only low quality evidence to support the theory that shaking an infant can cause such injuries[1]. This low quality evidence comes from confession based studies, which are hugely unreliable for a range of reasons[2] and are not scientific.


Regardless, everyone agrees that people should not shake babies, and that we need to be vigilant when it comes to child abuse. For this reason, it is generally agreed that Abusive Head Trauma (shaking, or shaking with impact against a soft surface that would not leave an outward trace) should be a differential diagnosis for infants who have the "triad" but who show no outward sign of abuse. So in such cases, police should check for witnesses that can prove that abuse caused those internal injuries.


So far, there is no real controversy and if we stopped here, ReDHI would be almost identical to SUID: in both cases doctors test for known medical causes, and if not found, they check for evidence of accidents. If neither is found, non-medical evidence of abuse may be explored, such as witnesses to a particular act of violence.


It is here that the diagnostic methodology for SUID and ReDHI split. If non-medical evidence of abuse is not found in SUIDs, the case is labeled “unknown” and called SIDS. By contrast, if non-medical evidence of abuse is not found for ReHDI, many doctors label the case as “Shaken Baby Syndrome” or Abusive Head Trauma (AHT).


This is not because a test has revealed that abuse occurred, or because there is non-medical evidence of abuse, such as witnesses. It is because abuse is used as the default diagnosis for the triad of injuries. When a cause cannot be found, abuse is diagnosed.


So in cases of the triad (or ReDHI), when there is no external evidence of trauma, abuse becomes the default differential diagnosis, even though there is only low quality evidence that it should be a differential diagnosis at all.


Many of the tests done in a differential diagnosis are not conclusive, only suggestive. These are very complicated conditions and diagnoses, and require a range of specialists. Sometimes the information required to diagnose possible medical conditions is just not available, whilst some tests are very expensive and time consuming. Many times, tests just lower the probability of a specific medical cause without totally ruling it out. Yet when a cause cannot be identified, rather than writing that a cause has not been identified and remains “unknown” (as occurs with SIDS), doctors time and again label these cases as abuse.


Those making the diagnosis of abuse often claim that the triad was not used to diagnose abuse, because abuse was only diagnosed after a differential diagnosis. This claim is specious: the only (supposed) evidence that abuse occurred is the triad. Abuse is the default diagnosis for the triad.


The experts do often claim that there is some feature of the triad that signifies abuse, such as the pattern or extent of retinal haemorrhages. But there is no valid scientific evidence that any manifestation of the triad can accurately diagnose abuse. For example there is no evidence[3] that extensive or particular patterns of retinal haemorrhage are diagnostic of abuse. Yet abuse is diagnosed in these cases time and again, and this has been going on for decades.


Allow me to complete my analogy with SIDS. For infants presenting with ReDHI, the reasoning is:

  1. we know that trauma is one way that these types of injuries can develop

  2. we are not able to identify a known medical cause with the tests we have made

  3. so, by default, it must have been trauma.

  4. but we cannot identify that accidental trauma occurred

  5. therefore, by default, it must have been abusive trauma.


So abusive head trauma is diagnosed, even though there are no external signs of injury and no evidence that abuse occurred. There are just an unexplained triad of internal injuries.


One could follow exactly the same logic in SIDs cases, simply by substituting the word suffocation for the word trauma


  1. we know that suffocation is one way that these types of injuries can develop

  2. we are not able to identify a known medical cause with the tests we have made

  3. so, by default, it must have been suffocation.

  4. but we cannot identify that accidental suffocation occurred

  5. therefore, by default, it must have been abusive suffocation.


Imagine if this did occur with SIDS and we routinely put one of the grieving parents in jail? Would that not seem outrageous?


Yet this line of reasoning is followed routinely with “Shaken Baby Syndrome”. In fact, there is far stronger evidence that suffocation should be a differential diagnosis for SUIDS than abuse should be a differential diagnosis for ReDHI. So why is abuse the default diagnosis when infants present with the triad? Why don’t these cases simply get labeled as “unknown” and called retino Dural Haemorrhage in Infant Syndrome, ReDHI?


If you are looking for a logical answer to that last question, you clearly do not know anything about Shaken Baby Syndrome[4]. It is not based on logic, on science or on rational reasoning.


In our criminal justice system, we (supposedly) convict people when there is evidence that they committed a crime. Doing a differential diagnosis, sometimes called a “work up” by the doctors, does not provide evidence for abuse. A test for abuse does not exist. A differential diagnosis simply looks for known medical causes, which can be missed by tests, and cannot account for unknown causes. If no known diagnosis is found, one cannot then use the triad, or any manifestation of the triad, as evidence of abuse: the triad does not amount to evidence from which abuse can be accurately diagnosed.


It is time doctors recognised the limitations of medical knowledge (as they do with SIDS) and took the need for an evidence-base seriously (as they do with SIDS). Most importantly it is time for doctors to show the same empathy for parents of infants who tragically die with symptoms of ReDHI, as they show to parents of infants who tragically die of SIDS.


If there is no evidence of abuse, the default must be to support grieving parents whose children die of unknown causes. Not to accuse them and to lock them up.



[1] Elinder et al. 2018 [2] E.g. Brook 2019 [3] Assuming that one does not count circular reasoning as evidence, a reasonable assumption for most people. And yes, this criticism applies to studies such as Bhardwaj et al. 2017 which purport to avoid the methodological flaws of other studies, but do not. Such studies lack a proper reference standard. [4] or abusive head trauma or whatever you want to call these cases when there is no external sign of injury

229 views3 comments

Recent Posts

See All

Why is Shaken Baby Syndrome never witnessed?

In a café in Carrum Downs in 2016, a frustrated 39 year father angrily shook his 5 month old baby. Concerned onlookers called police and the man was charged and convicted for assault. Thankfully, his