pancreatic cancer increases with higher blood sugar, even if prediabetic

 A recent study found that cumulative hyperglycemic exposure was associated with increased risk of pancreatic cancer (see dm hyperglyemic burden inc pancreat cancer JDiabResClinPrac2022, or  doi.org/10.1016/j.diabres.2022.110208 )

Details:

-- 3,138,099 individuals who had 4 consecutive annual health screenings between 2009 and 2013, from the Korean National Health Insurance Service Database of Claims

    -- this database of claims data has medical information of 50 million Koreans (>99% of the population), including age, sex, primary and secondary diagnoses, hospital visits, prescriptions (inpatient and outpatient).

-- age 45, 75% male, 33% smokers, 10% heavy drinkers, 22% regular exercise, 16% low income, 35% obese (range 27%-49%), 30% hypertensive (12%-47%), 25% dyslipidemia (12%-35%), 2% cardiovasc disease, 35% metabolic syndrome (9%-63%), BMI 24, fasting glucose 110mg/dL (87-134), BP 124/80

    -- there was a strong trend to worse demographics/medical issues with increasing hyperglycemia burden. the ones with ranges above had the most extreme differences comparing lowest to highest

-- hyperglycemic burden was defined as the cumulative score of blood glucose:

    -- a score of 0-4 got 1 point for each of the 4 times that the blood glucose was >100 mg/dL or the patient was on an antidiabetic drug, for a score from 0-4

    -- also, a more quantitative assessment: 1 point for blood glucose 100-125 mg/dL, and 2 points if >125 mg/dL for each of the 4 years, leading to a total score in the 0-8 range

-- main outcomes: risk of pancreatic cancer by hyperglycemic burden and by hyperglycemic burden score

-- follow-up for 6.2 years

 

Results:

-- risk of pancreatic cancer by hyperglycemia burden (0-4):

    -- 0: 1,424,156 people (45%), the reference for below, all for fully adjusted model

    -- 1: 741,330 people (24%), 15% increase, HR 1.15 (1.04-1.28)

    -- 2: 416,425 people (13%), 30% increase, HR 1.30 (1.16-1.45)

    -- 3: 260,536 people (8%), 26% increase, HR 1.26 (1.12-1.43)

    -- 4: 295,625 people (9%), 67% increase, HR 1.67 (1.51-1.85)

        -- this trend was highly statistically significant, in their unadjusted and 2 adjusted models, p<0.0001

-- risk of pancreatic cancer by hyperglycemia burden score (0-8):

    -- 0: 1,424,156 people, the reference for below, all for fully adjusted model

    -- 1: 705,119 people, 14% increase, HR 1.14 (1.03-1.27)

    -- 2: 407,865 people, 30% increase, HR 1.30 (1.16-1.45)

    -- 3: 243,441 people, 23% increase, HR 1.23 (1.04-1.40)

    -- 4: 143,852 people, 35% increase, HR 1.35 (1.17-1.57)

    -- 5: 50,273 people, 59% increase, HR 1.59 (1.30-1.94)

    -- 6: 30,249 people, 66% increase, HR 1.66 (1.30-2.11)

    -- 7: 24,704 people, 69% increase, HR 1.69 (1.30-2.19)

    -- 8: 108,440 people, 89% increase, HR 1.89 (1.66-2.14)

      -- this trend was highly statistically significant, in their unadjusted and 2 adjusted models, p<0.0001

-- comparing the whole population to the non-diabetic ones (diabetes defined as at least one claim per year for a prescription anti-diabetic med, or glucose >126 mg/dL):

-- patients who had a pre-diabetes glucose level on at least one occasion had an overall 14% increased risk of pancreatic cancer vs those with consistent normoglycemia, increasing gradually with hyperglycemic burden score (those with score of 8 had an 89% increased risk)

Commentary:

-- diabetes is known to be associated with several cancers: breast, ovarian, GI, hepatocellular, and it seems with pancreatic cancer; estimates are that all cancers are increased 1% in diabetics.

-- pancreatic cancer is globally a leading cause of death. In the US, the SEER registry estimated that the number of new cases in 2022 would be 62,210 (3.2% of all new cancers), the estimated deaths in 2022 would be 49,830 (8.2% of all cancer deaths). And the 5-year survival from 2012-2018 was found to be only 11.5%: see https://seer.cancer.gov/statfacts/html/pancreas.html and https://academic.oup.com/ije/article/47/2/427/4628151 

    -- for a blog on the increasing pancreatic cancer rate in the US, and some possible explanations, see http://gmodestmedblogs.blogspot.com/2021/12/pancreatic-cancer-increasing-incidence.html .

-- this current study adds to other cohort studies suggesting that blood glucose levels are related to pancreatic cancer, though this study goes beyond those assessing only a baseline blood sugar, since it incorporates 4 measurements over 4 years. and, as we know, things do change: some people have increasing blood sugars and some decreasing over time, largely associated with changes in their lifestyles (diet, exercise, weight, etc)

    -- several studies based on single blood sugar measurements at baseline have found the same results: both pre-diabetes and diabetes are associated with increased pancreatic cancer

-- another advantage of this Korean study is it makes reverse causation less likely (ie, pancreatic cancer damaging the pancreas and thereby leading to diabetes). with 4 annual visits for patient information and a 6.2 y follow-up, this study makes reverse causation less likely (at baseline, they had diagnosed pancreatic cancer in 0.06% in those with no hyperglycemic burden but up to 0.25% in those with burden of 4. but to be certain of no reverse causation, there would need to be (an unthinkable) randomized controlled trial where patients were randomized and assigned to different levels of blood sugar levels and followed for years...

 

-- this current study, finding increased pancreatic cancer in patients who are prediabetic, really brings up the issues of how we define "diabetes":

    -- people with prediabetes are at a significantly higher risk for cardiovascular disease.  several studies have found that an A1c even down to 5.7% (some even lower than that, especially in men) seems to be associated with significant cardiovascular disease and mortality, in the 25-30% range (see http://gmodestmedblogs.blogspot.com/2016/12/prediabetes-and-cardiovascular-risk.html). 

    -- the baseline concern here is that our definition of diabetes by A1c of >6.5% is, I believe, mistargeted: the cut point of 6.5% reflects the point of inflection when diabetic retinopathy increases. And, though retinopathy is certainly not a good thing, it is not what is highly fatal (80% of diabetics die from cardiovascular causes). Our current definition leads to the misperception that all is good if the A1c is <6.5%

        -- ie, the whole definition of prediabetes seems to be based on a largely non-pathophysiologic cutpoint that obscures the importance of lower levels of blood sugar for cardiovascular disease and pancreatic cancer:  the glucose levels in these (and perhaps other) diseases is a continuum that is obscured by this cutpoint

    -- and, this issue of “pre-diabetes” takes on increasing significance as magnitude of the overall levels of overweight/obesity are increasing so dramatically and the numbers of kids and adults with this label is increasingly hugely

    -- the number of undiagnosed diabetes is also pretty staggering: a study from 2014 found that 46% (175 million) of diabetic patients were undiagnosed, with 24%-75% across the globe (higher numbers in low- and middle-income countries): https://www.sciencedirect.com/science/article/abs/pii/S0168822713003847

        -- since diabetes (type 2) is largely asymptomatic for years (and accumulating morbidities: up to 1/3 of people in some studies already have microvascular changes at the time of diagnosis), this all strongly suggests that we should be doing pretty aggressive diabetes screening on a regular basis (A1c, but also eye exams and other physical exams to detect early microvascular changes), and being really aggressive in our treatments (diet/exercise/weight management, as well as more aggressive lipid lowering). Studies have (not surprisingly) shown that undiagnosed diabetes has a similar 1.5- to 3-fold higher mortality risk.

        -- based on many factors, the American Diabetes Association in their current 2023 recommendations (see https://diabetesjournals.org/care/article/46/Supplement_1/S19/148056/2-Classification-and-Diagnosis-of-Diabetes ) recommends screening: those with first degree relatives with type 2 diabetes, those from "high risk" ethnicities (African American, Latino, Native American, Asian American, Pacific Islander), history of CVD, blood pressure >130/80 or on meds, HDL <35 or triglycerides >250, those with polycystic ovary syndrome, physical inactivity, those with prediabetes/gestational diabetes, and anyone else who is at least 35 years old. [ie, approximately 100% of the adult population, it appears......]

    -- but, bottom line, it behooves us clinicians to be attentive to patients with lower A1c levels, perhaps arbitrarily in the 5.5-5.7% range, and treat these people pretty aggressively: strong and repeated reinforcement of diet and exercise, and (perhaps) more aggressive use of statins. and maybe even lower thresholds for starting metformin....

Limitations:

-- Problems with this database: no information on actual causes of death. the population limited to one country with different social/demographics/cultural/medical issues vs others (eg, in this study, a relatively healthy group with little hypertension, low BMI, rare cardiovascular disease, etc, and all from Korea)

-- this was a quick and dirty data-mining study, missing the nuances of a more accurate dose-response curve, needing more granular data (not just their 2 buckets of blood sugar 100-125 or >125 mg/dL) to see the strength of the relationship with pancreatic cancer, as well as if the relationship is linear vs plateauing (vs decreasing??) with higher cumulative blood glucose burden. and their scoring system is not one that has been validated

-- not lots of granular data on other issues: they did not stratify individuals in being on meds vs not (or even if they started or stopped meds during the study), if exercise/dietary changed over time (with fine gradations of amounts of change), if there were changes in sociodemographics over time, changes in medical issues over time (other cancers, other conditions that might influence results, changes in medications that could affect results....), 

    -- they give a high score for glucose burden on all people on antidiabetic meds. But, especially with the newer agents, many diabetics have blood sugars in the normal range. Are they still at higher pancreatic cancer risk?

-- the definition of diabetes vs nondiabetes, including being on meds, was imperfect, since many "pre-diabetics" may be put on a diabetes med (eg metformin) to prevent development of "diabetes"

So,

-- diabetes is associated with several cancers, with pancreatic cancer likely to make it onto the list

-- though it is not clear from this study, it does add to the imperative to diagnose diabetes (and pre-diabetes) early and treat it aggressively (nonpharmacologically and pharmacologically if needed), as recommended by several organizations (eg American Diabetes Association, with their 2023 recommendations of screening noted above). again, no great data on this, given the nature of observational studies and the lack of ability to set up an RCT with a large cohort of those with diabetes being treated and another not

-- and, this issue of diabetes and impaired glucose tolerance is such a huge and increasing public health issue, far outweighing the issue of pancreatic cancer (though a pretty miserable cancer) in terms of personal, family, community, and national costs (with "costs" being used broadly in terms of personal and social disruptions as well as personal and social financial costs)

-- we probably should also more aggressive in explaining to patients and treating them more aggressively by promoting lifestyle changes and perhaps earlier introduction of meds when they are in the "pre-daibetes" range....

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

  

to get access to all of the blogs (2 options):

1. go to https://www.bucommunitymedicine.org/ , a website from the Community Medicine section at Boston Medical Center.  This site does have a very searchable and accessible list of my blogs and is the easiest to view blogs and displays more at a time.

2. go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

  -- click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

  -- or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them

  

if you would like to see the article, please email me. 

please feel free to circulate this to others. also, if you send me their emails, i can add them to the list