According to a recent report, type 2 diabetes (T2D) caused the deaths of over 100,000 Americans in 2021, the second year in a row that diabetes-related deaths have reached this mark. These numbers are certainly alarming, and many have pointed out the steep jump from the approximately 87,600 diabetes-attributed deaths in 2019. Still, even in 2020, the death toll from diabetes was only a fraction of that from the two leading causes, heart disease (~700,000) and cancer (~600,000). The CDC reports that COVID-19 accounted for 3.5x as many deaths in 2020 as T2D, and mortality associated with accidents or unintentional injury outpaced diabetes mortality by about 2:1. In fact, diabetes was only the 8th-leading cause of death in 2020, ranking just above influenza and pneumonia.
Can it really be possible that diabetes is so relatively… harmless? Of course not. So why don’t the mortality data reflect the severity of the disease?
Diseases do not exist in isolation
The truth is that, more often than not, death cannot be attributed to one specific, isolated cause, and casting each case into a neatly-defined category rarely tells the full story. Each of the most deadly diseases in the developed world do not exist in isolation; one condition can influence the risk of developing another, and multiple conditions can exist simultaneously and interact in ways that critically impact mortality risk. (To illustrate this point, consider that Freddie Mercury’s immediate, official cause of death was bronchial pneumonia, though this explanation on its own certainly fails to provide an accurate account of the true, underlying cause.)
The complexity and uncertainty surrounding causes of death and their categorization has resulted in discrepancies for certain diseases between their true impact on mortality and the impact as reported by the CDC. This point came up recently in one of my COVID newsletters, in which I mentioned that deaths attributed to the Omicron variant have likely been inflated by an unknown percentage representing individuals who died of other causes after incidentally testing positive for COVID. As I indicated in that discussion, the most accurate way to quantify the true mortality attributable to a given cause is to determine how many fewer deaths would have occurred if that cause weren’t present. Estimating this difference may not be trivial, but in the case of diabetes, the directionality is clear: if T2D didn’t exist, is it possible that “only” 100,000 fewer deaths would occur? Not a chance.
Patients with T2D have a significantly elevated risk of cardiovascular disease (CVD), cancer, and Alzheimer’s disease (AD) relative to those without T2D, yet deaths from those respective diseases count as their own categories. In addition, T2D and its characteristic hyperglycemia can increase the severity, progression, and mortality risk of these other diseases. For example, one cohort study of colon cancer patients found that over a median 9.4-year follow-up, diabetic patients had a 21% increased risk of cancer recurrence and a 42% increased risk of all-cause mortality relative to non-diabetics after correcting for confounding factors, including body mass index (BMI). So even if, say, cancer were to develop regardless of diabetes status, the presence of T2D raises the likelihood that the cancer will prove fatal. To put it another way, for many cases in which cancer is the immediate or primary cause of death, the true cause is the combined effect of diabetes and cancer, though only the latter is reported.
Why don’t we report all contributors to death?
So why doesn’t the CDC report all contributors to death for cases in which multiple diseases are at play? Simply put: because they can’t. On an individual level, it can be nearly impossible to ascertain the relative contributions of myriad comorbidities. On a population level, statistical analyses may be employed to shed light on approximate risk ratios associated with other conditions, but even these calculations are fraught with confounding variables and questions about causation. Even if it were possible to determine precisely every contributing factor to every death, how would it be reported? Listing a single death under multiple broad categories, resulting in a grossly overestimated number of total deaths? Or creating so many highly-individualized categories (cancer, cancer + T2D, cancer + AD, cancer + T2D + AD, etc…) that we lose any utility of categorization in the first place?
In short, the CDC reports isolated causes of death in broad categories because it has no other choice – not because this strategy is necessarily an accurate representation of actual mortality associated with every disease. There is no doubt that T2D is among those diseases for which deaths are drastically underreported. So although 100,000 Americans is a frightening number, we must bear in mind that a much larger impact on mortality lies hidden, and the CDC reports are only the tip of the iceberg.