Dead in Bed Syndrome   (July 29, 2013)

Twitter recently got in a twit about a statement someone found on the Internet: “One in 20 type 1 diabetics die in their sleep due to a sudden drop in their blood sugar.” That’s quite a statistic, and one that raises all sorts of questions. Where did this strange and disturbing statistic come from? Does this mean that of all T1 PWD who die, 1 in 20 (5%) die in their sleep from something relating to dropping blood sugar? Does “sudden drop” mean any kind of drop – for instance, from high to normal, or only if it goes low? What’s the evidence that these deaths are indeed due to low blood sugar? And what can we do to prevent such a death?

The “Dead in Bed Syndrome” is quite a problem, both for parents of kids with diabetes, and their physicians. One pediatric endocrinologist said “my patients are totally freaked out about this (as am I). My problem is that we have about 1200 patients in our practice with type 1 dm- does that mean statistically 6 patients in my practice will eventually drop dead in their sleep.”

A thorough discussion with multiple references, is on-line at the Children With Diabetes website, at The Dead in Bed Syndrome. It should be noted that different authors have developed different definitions for the Dead in Bed Syndrome (sometimes abbreviated DIB), and partially as a result of the differing definitions, and probably mainly as a result of differing patient populations (e.g., country where the study was done, degree of diabetes control of the patients, age breakdown, etc.) the rate of DIB varies widely.

The Dead in Bed Syndrome was first discussed in 1991 when the Professional Advisory Committee of the British Diabetic Association published a report, Unexplained deaths of type 1 diabetic patients . They evaluated  50 autopsied deaths of people with Type 1 diabetes under age 50 years in the UK in 1989. They found that “22, aged 12-43 years, most of whom had gone to bed in apparently good health and been found dead in the morning: Nineteen of the 22 were sleeping alone at the time of death and 20 were found lying in an undisturbed bed. Most had uncomplicated diabetes and in none were anatomical lesions found at autopsy… the timing of death and other circumstantial evidence suggests that hypoglycaemia or a hypoglycaemia-associated event was responsible.” That’s 22 cases out of 50 deaths, or 44%, but the fifty deaths were clearly selected cases.

A somewhat more recent article (1999), Dead-in-bed syndrome in young diabetic patients stated that “It appears that such deaths occur in 6% of all deaths in diabetic patients below age 40 years.” Again, the authors were concerned that hypoglycemia was involved: “The causes are by definition unknown, but a plausible theory is a death in hypoglycemia, since a history of nocturnal hypoglycemia is noted in most cases.”

At about the same time, a literature review concluded that the syndrome was somehow related to both neuropathy and heart rhythm problems: “the 'dead in bed' syndrome probably occurs in Type 1 diabetic persons with early autonomic neuropathy, resulting in relative sympathetic overactivity. In such persons, risks of ventricular dysrhythmias will be compounded by nocturnal hypoglycaemia, which may be associated with an increase in the electrocardiographic Q-T interval, and Q-T dispersion. This could lead to the observed sudden death in undisturbed beds.” In other words, low sugar in PWD who had autonomic neuropathy might have led to a heart rhythm disturbance that caused the death.

In 2008, another article stated that  the dead-in-bed syndrome “accounts for 5-6% of mortality cases in patients with type 1 diabetes.” No definition of the syndrome, nor age cutoff, were provided in the abstract for this article.

In 2011, a review of two diabetes registries found that seven of 329 (only 2%) of participants who had died met their strictly-defined criteria for dead-in-bed syndrome: sudden unexplained death  in otherwise healthy type 1 patients,  age less than 50, who were found dead in bed. The cause of death in the seven dead-in-bed persons was ascertained  as: diabetic coma (n=4), unknown (n=2) and undiagnosed heart disease (n=1, cardiomyopathy found on autopsy).

Then in 2011 someone at the JDRF got a bright idea, and made everyone even more anxious: in an advertisement in November 2011 that they ran in both the New York Times and the Washington Post, they presented a photograph of a smiling little girl (wearing a medical ID bracelet) as being “Piper, age 8, diagnosed at age 3” and said that “Piper has type 1 diabetes. One in twenty people like Piper will die from low blood sugar.” Note, there is nothing in the advertisement, and very little in subsequent discussions of the ad, to imply that Piper would specifically die of the DIB syndrome. Although it’s sad that it can happen, PWD do occasionally die of hypoglycemia; I’ve discussed it before, at Bad Outcomes From Hypoglycemic Coma I don’t know if the correct rate for hypoglycemic deaths is 1 in 20 – but from my memories of my diabetes practice, of mainly adult PWD, I wonder if that’s an accurate number.

But the problem I’m discussing in this essay, namely the sad statistic that some uncertain percentage of people with T1DM will die in their sleep, isn’t the same issue that the JDRF chose to publicize.

After all, it’s not even proven that the DIB syndrome is caused by hypoglycemia. And there’s nothing I’ve read that says that death might be caused by falling blood glucose levels that don't drop into the hypoglycemic range: for instance, that someone who’s blood sugar was 400 at bedtime and dropped, let’s say, to 150, would be at risk of DIB syndrome. And I can’t find anything to say that younger (or older) PWD are more “at risk” – many of the scientific discussions use an upper-age cutoff (such as 40 or 50 years of age) so the risk of elderly PWD of dying of DIB syndrome isn’t at all clear.

So I don’t think the statement with which I opened this essay is correct that “One in 20 type 1 diabetics die in their sleep due to a sudden drop in their blood sugar”—but it very well might be that “approximately 1 in 20 apparently healthy younger people (under about age 40 or 50) who have T1DM die in their sleep from unknown causes that might be hypoglycemia-related.”

Assuming my modification of the statement might be true, then the next question is whether  DIB could be preventable. Would checking a blood glucose at 3AM (or checking multiple times every night, depending on how compulsive the patient or caregiver might be) be sufficient to see if hypoglycemia is present, and would treatment of the low sugar prevent a catastrophe? That was all that we had available until a few years ago.

Nowadays, there’s a possibly better answer: anyone who is on insulin, or who is at risk of hypoglycemia for other reasons, should have a continuous glucose monitoring (CGM) device with a loud alarm that triggers when the glucose level is falling rapidly, or when the glucose level falls below an arbitrary pre-set level. (I can’t count how many times my CGM has alerted me, or my wife, or both of us, to falling levels at night!)

Assuming a causative role of hypoglycemia in the Dead in Bed Syndrome, routine use of CGM should decrease -- and perhaps eliminate -- the risk of DIB. If you, or your loved one, are taking insulin, and have ever had hypoglycemia at night, you probably should be using CGM. It just might be a life-saver.