From the list of 10 diseases that 23andMe reports on, three were listed in the Elevated Risk category. We have already posted on Crohn’s disease and Prostate cancer. The remaining one is Type I Diabetes. I am lucky that I do not suffer from any of these conditions, but I value the opportunity this experience has given me to learn about them, what my risks might be, what can I do about it, and what should I discuss with my doctor if symptoms arise.
Click Image to Enlarge
Last Updated
From the graphic above, did you notice that my information on Type I Diabetes has recently changed? The date under Last Updated is July 30, 2009. The report tells me that the update is because “the risks associated with SNPs rs1990760, rs1893217, rs2476601, rs3184504 and rs725613 were updated with newer values. As a result, some customers may find that their results in the Odds Calculator differ slightly from previous values.”
When I looked back, my original lifetime scores for Type I Diabetes were Absolute Risk 1.9% (now 1.7%), and Relative Risk 1.91 (now 1.67). I must be living right.
Comparing Type I with Type II
As I learned with the other diseases, my risk changes dramatically when I use the Odds Calculator and enter my age range. Although still higher than the average, my risk of developing Type I Diabetes is only 18 out of 10,000. The calculator told me that my biggest risk would have been when I was 10-14 years old.
Click Image to Enlarge
When I look at my result for Type II Diabetes, I find it under the Typical Risk category. But the odds for me developing Type II are 130 out of 10,000, or more than 7 times that of Type I. Yet Type I is in the Elevated category. My first thought was why would my risks be so different between Type I and Type II?
I understand that these risk categories are comparing me to the general population (Relative Risk). But if I ranked the 10 diseases by Absolute Risk, they would fall into different categories I would create, like the Greatest Threat category (which would include Prostate cancer 24%, Type II Diabetes 19%, Venous Thromboembolism 12%), followed by the Moderate Concern category (Psoriasis 9.9%, Age-related macular degeneration 4.3%, Rheumatoid arthritis 2.3%), and finally the Least of My Worries category (Type I Diabetes 1.7%, Crohn’s disease 1.6%, Parkinson’s disease 1.6%, and Celiac disease 0.04%). Of course this ranking could change based on any family history I discover.
More Heritability than the Other Diseases
Although the suspected environmental factors for Type I Diabetes are interesting (living in a cold climate, not being breast fed as an infant), out of the ten 23andMe reported diseases, Type I has the highest incidence with 72-88% attributable to genetics.
Click Image To Enlarge
In second place is Celiac diseases with 57-87%. Psoriasis with 66-80% came in third. At the bottom is Parkinson’s disease with only 0-1% - which I don’t understand how that could be so low. Type II Diabetes was second to last with 26%, again quite a bit different from Type I.
8 Markers for Type I
I find myself looking more closely at the markers reported on for each disease. Crohn’s had 12. Prostate had 5. Type I Diabetes is in the middle of those two with 8. Type II is close with 9. The other reported diseases have only one or two reported markers. Why the broad range in the number of markers? This is another bit of evidence that trying to figure this out is complicated.
Click Image To Enlarge
The SNP reported here in the HLA region had the strongest association with Type 1 Diabetes in one of the largest studies done so far (see the reference below). 23andMe then made an interesting alteration with this note –
The 2007 Wellcome Trust paper reported a strong association between type 1 diabetes and the SNP rs9272346. Our quality control process flagged data for rs9272346 as unreliable, so we instead included the SNP rs3129934, which is also highly associated with type 1 diabetes. Both are in the HLA region, although the exact linkage patterns between tagging SNPs and traditionally determined HLA haplotypes is still being worked out.
Click Image To Enlarge
No Family History
I can’t think of any 1st, 2nd, or even a 3rd degree relative that has diabetes of any kind. This is good because apparently there is not much one can do try to prevent the onset of Type 1 Diabetes. Here is something I learned from the MD’s Perspective section- a Q&A with Dr. Roshanak Monzavi, M.D., Assistant Professor, Center for Diabetes, Endocrinology and Metabolism, at Children's Hospital of Los Angeles -
23andMe: Historically, how have doctors used genetics or family history when assessing a person's risk of developing Type 1 Diabetes (T1DM)?
Dr. Monzavi: Unfortunately, there is currently no proven preventive measure that can be used to decrease the risk of development of T1DM in people with family history of T1DM. However, physicians are studying some interventions that may prevent T1DM or postpone its time of presentation through clinical trials such as TrialNet. Family members of people with T1DM are evaluated for certain genetic and/or immunologic markers, which put them at high risk for T1DM, and if at high risk, they may enroll in such a study to assess the efficacy of an intervention in preventing T1DM.
I discovered that a TrialNet study tests whether a daily oral insulin capsule can prevent or delay the disease in at-risk relatives of people with Type 1 Diabetes. TrialNet is also performing a pilot study testing whether docosahexaenoic acid (DHA), and omega-3 fatty acid found in some foods can prevent or delay the autoimmunity leading to type 1 diabetes. Let’s hope they succeed.
Perspective from a medical geneticist
Communicating Risk
One of the real struggles we have in genetics is trying to explain probabilities and risks to our patients. Part of this relates to our brains operating more as pattern recognition engines than analytic computers. The second and more important issue is that there really isn’t a ‘right’ answer to the risk perception question, unless one considers the individual’s perspective. We see this in the prenatal clinical setting all the time.
For some couples a slight increase in risk for a certain birth defect (say a 5% risk of having another child with a congenital heart defect) is perceived as so high as to necessitate a battery of testing options, or in some cases choosing not to have more children. In contrast, other couples at a 50% risk for an anomaly will have very few concerns about subsequent pregnancies. Thus we must expand beyond absolute and relative risks to what might be characterized as personal perception of risk.
Aligned With Absolute Risk
In his post, Grant developed his own risk categories: Greatest Threat; Moderate Concern; and Least of My Worries. He then categorizes his diseases by absolute risk such that the highest absolute risks are in Greatest Threat (with an apparent absolute risk cutoff of 10%) and the lowest are in the Least of My Worries (with an apparent cutoff of 2%). As with previous posts, Grant has clearly aligned himself with Absolute Risk. He acknowledges that discovery of relevant family history could change the ranking (it appears daily association with geneticists has successfully brainwashed him). But how would it change the assignment? Would Grant still rank this according to an adjusted absolute risk or would family history of disease have a privileged status in ranking risk?
What If the Disease Is Preventable
Another potential factor to consider would be the availability of measures to prevent disease. Would this change Grant’s ranking of the diseases? We really can’t tell with the 10 reported diseases because only one of them (type II diabetes) has preventive interventions that seem to work—at least if one extrapolates from studies of hyperglycemic patients without diabetes. For Grant, this is a moot point, as he is at an appropriate weight, eats healthily and exercises. Frankly, based on lifestyle I’m at much higher risk than Grant, although I am also blessed with a completely negative family history of either type I or type II diabetes.
So why does this matter? One of the answers has to do with what could be characterized as the utility of the information. In the MD Perspective for both type I and type II diabetes on the 23andme website both physicians acknowledge that there is currently no way to use the genomic predisposition to prevent diabetes. From the medical perspective the testing has no utility, in that the results of the test do not lead to alterations in care that improve the patient’s health outcome. This type of information is frequently used by insurers or health systems to decide whether or not a given test should be covered or even offered to patients.
Personal Utility
There is another concept that is beginning to emerge in discussion of genomic testing which is the idea of personal utility. What will this test result mean to me as a person? What will I do if the result is positive or negative? If there is no insurance coverage what am I willing to pay out of my own pocket because I believe this will help me? This principle clearly operates in medicine all the time. The supplement and nutriceutical industry not only survives but thrives as a multi-billion dollar industry almost solely on the basis of personal utility, given that there is limited evidence of increased health benefit (and despite several instances of overt harm. Fen-Phen anyone?)
Folate Flashback
I took a daily Folate supplement for a couple of years because I knew: 1) I didn’t eat as many green vegetables as I should; 2) Folate decreases serum homocysteine levels; 3) elevated homocysteine is associated with an increased risk of cardiovascular disease; 4) Even if it doesn’t help, it surely won’t hurt (and it’s easier than eating vegetables!!); 5) It’s cheap. As Folate was systematically studied, it was found that for reasons that still are unexplained, men who had higher Folate intake had more cardiovascular events. Into the trash with my Folate pills!!
Bottom line was my decision to spend discretionary income on Folate was based on a belief that Folate supplements would reduce an outcome of concern for me (a heart attack) at an acceptable ‘cost’ (price, lifestyle, added bother to remember to take the pill, etc.). For me, the ‘cost’ of actually losing weight, eating more healthily, exercising regularly, which does have demonstrated utility, is too high. My perception of the personal utility of these health behaviors is too low to justify spending time and effort to achieve them, despite abundance evidence to the contrary.
The Patient Perception Point
So what’s the point? (Aha, you assume I have one). From my perspective the issue is that we frequently ignore the disconnect between evidence of utility and a patient’s perception of utility. This leads to recommendations based on sound evidence falling on apparently deaf ears. If we make no attempt to understand what the patient is interested in and why they are interested in it, we may miss opportunities to improve health in ways we don’t anticipate. I now know that Grant is most worried about Prostate Cancer, Type II diabetes and Venous Thromboembolism. Let’s assume that Grant’s health behaviors aren’t as good as they should be. If he were overweight, would his concern about Type II diabetes based on his 23andme results present a teachable moment about what he could do to control his risk?
No way to know for sure, but it’s more likely than if his major worry was Crohns disease. I might say, Gee, Grant, you’re really concerned about diabetes and blood clots. Did you know that being overweight really increases your risk for both of those problems? In fact, losing weight has been proven to dramatically reduce your risk of developing diabetes. Is this something you’re ready take on, because if you are, I have a dietician that can help you assess your diet, and I have a list of weight loss programs that will help you to not only lose the weight, but will help you to keep it off.
The bottom line for me is listen to the patient, read between the lines, get to the patient’s concern and help them to develop a plan to address the concern. If it takes a 23andme test to get to the concern for that patient, I can live with that—and maybe they can too!
And now for something completely different
This is a complete non-sequitur, but as I was playing in Grant’s 23andme site (he lets me log in for the purposes of the blog and I only look at the topic under discussion) I found a tab called timeline. Clicking on this gives a timeline of major discoveries for the disease of interest. In the case of diabetes this goes back to 1550 BC to a citation in the Papyrus Ebers by an Egyptian physician Hesy-Ra describing sweet urine and a variety of dietary cures and ends in 2007 with the Wellcome Trust case control genome-wide association study for type I diabetes. I’m showing my geekitude (geekosity?, geekness?), but I thought it was cool.
Click Image to Enlarge
Last Updated
From the graphic above, did you notice that my information on Type I Diabetes has recently changed? The date under Last Updated is July 30, 2009. The report tells me that the update is because “the risks associated with SNPs rs1990760, rs1893217, rs2476601, rs3184504 and rs725613 were updated with newer values. As a result, some customers may find that their results in the Odds Calculator differ slightly from previous values.”
When I looked back, my original lifetime scores for Type I Diabetes were Absolute Risk 1.9% (now 1.7%), and Relative Risk 1.91 (now 1.67). I must be living right.
Comparing Type I with Type II
As I learned with the other diseases, my risk changes dramatically when I use the Odds Calculator and enter my age range. Although still higher than the average, my risk of developing Type I Diabetes is only 18 out of 10,000. The calculator told me that my biggest risk would have been when I was 10-14 years old.
Click Image to Enlarge
When I look at my result for Type II Diabetes, I find it under the Typical Risk category. But the odds for me developing Type II are 130 out of 10,000, or more than 7 times that of Type I. Yet Type I is in the Elevated category. My first thought was why would my risks be so different between Type I and Type II?
I understand that these risk categories are comparing me to the general population (Relative Risk). But if I ranked the 10 diseases by Absolute Risk, they would fall into different categories I would create, like the Greatest Threat category (which would include Prostate cancer 24%, Type II Diabetes 19%, Venous Thromboembolism 12%), followed by the Moderate Concern category (Psoriasis 9.9%, Age-related macular degeneration 4.3%, Rheumatoid arthritis 2.3%), and finally the Least of My Worries category (Type I Diabetes 1.7%, Crohn’s disease 1.6%, Parkinson’s disease 1.6%, and Celiac disease 0.04%). Of course this ranking could change based on any family history I discover.
More Heritability than the Other Diseases
Although the suspected environmental factors for Type I Diabetes are interesting (living in a cold climate, not being breast fed as an infant), out of the ten 23andMe reported diseases, Type I has the highest incidence with 72-88% attributable to genetics.
Click Image To Enlarge
In second place is Celiac diseases with 57-87%. Psoriasis with 66-80% came in third. At the bottom is Parkinson’s disease with only 0-1% - which I don’t understand how that could be so low. Type II Diabetes was second to last with 26%, again quite a bit different from Type I.
8 Markers for Type I
I find myself looking more closely at the markers reported on for each disease. Crohn’s had 12. Prostate had 5. Type I Diabetes is in the middle of those two with 8. Type II is close with 9. The other reported diseases have only one or two reported markers. Why the broad range in the number of markers? This is another bit of evidence that trying to figure this out is complicated.
Click Image To Enlarge
The SNP reported here in the HLA region had the strongest association with Type 1 Diabetes in one of the largest studies done so far (see the reference below). 23andMe then made an interesting alteration with this note –
The 2007 Wellcome Trust paper reported a strong association between type 1 diabetes and the SNP rs9272346. Our quality control process flagged data for rs9272346 as unreliable, so we instead included the SNP rs3129934, which is also highly associated with type 1 diabetes. Both are in the HLA region, although the exact linkage patterns between tagging SNPs and traditionally determined HLA haplotypes is still being worked out.
Click Image To Enlarge
No Family History
I can’t think of any 1st, 2nd, or even a 3rd degree relative that has diabetes of any kind. This is good because apparently there is not much one can do try to prevent the onset of Type 1 Diabetes. Here is something I learned from the MD’s Perspective section- a Q&A with Dr. Roshanak Monzavi, M.D., Assistant Professor, Center for Diabetes, Endocrinology and Metabolism, at Children's Hospital of Los Angeles -
23andMe: Historically, how have doctors used genetics or family history when assessing a person's risk of developing Type 1 Diabetes (T1DM)?
Dr. Monzavi: Unfortunately, there is currently no proven preventive measure that can be used to decrease the risk of development of T1DM in people with family history of T1DM. However, physicians are studying some interventions that may prevent T1DM or postpone its time of presentation through clinical trials such as TrialNet. Family members of people with T1DM are evaluated for certain genetic and/or immunologic markers, which put them at high risk for T1DM, and if at high risk, they may enroll in such a study to assess the efficacy of an intervention in preventing T1DM.
I discovered that a TrialNet study tests whether a daily oral insulin capsule can prevent or delay the disease in at-risk relatives of people with Type 1 Diabetes. TrialNet is also performing a pilot study testing whether docosahexaenoic acid (DHA), and omega-3 fatty acid found in some foods can prevent or delay the autoimmunity leading to type 1 diabetes. Let’s hope they succeed.
Perspective from a medical geneticist
Communicating Risk
One of the real struggles we have in genetics is trying to explain probabilities and risks to our patients. Part of this relates to our brains operating more as pattern recognition engines than analytic computers. The second and more important issue is that there really isn’t a ‘right’ answer to the risk perception question, unless one considers the individual’s perspective. We see this in the prenatal clinical setting all the time.
For some couples a slight increase in risk for a certain birth defect (say a 5% risk of having another child with a congenital heart defect) is perceived as so high as to necessitate a battery of testing options, or in some cases choosing not to have more children. In contrast, other couples at a 50% risk for an anomaly will have very few concerns about subsequent pregnancies. Thus we must expand beyond absolute and relative risks to what might be characterized as personal perception of risk.
Aligned With Absolute Risk
In his post, Grant developed his own risk categories: Greatest Threat; Moderate Concern; and Least of My Worries. He then categorizes his diseases by absolute risk such that the highest absolute risks are in Greatest Threat (with an apparent absolute risk cutoff of 10%) and the lowest are in the Least of My Worries (with an apparent cutoff of 2%). As with previous posts, Grant has clearly aligned himself with Absolute Risk. He acknowledges that discovery of relevant family history could change the ranking (it appears daily association with geneticists has successfully brainwashed him). But how would it change the assignment? Would Grant still rank this according to an adjusted absolute risk or would family history of disease have a privileged status in ranking risk?
What If the Disease Is Preventable
Another potential factor to consider would be the availability of measures to prevent disease. Would this change Grant’s ranking of the diseases? We really can’t tell with the 10 reported diseases because only one of them (type II diabetes) has preventive interventions that seem to work—at least if one extrapolates from studies of hyperglycemic patients without diabetes. For Grant, this is a moot point, as he is at an appropriate weight, eats healthily and exercises. Frankly, based on lifestyle I’m at much higher risk than Grant, although I am also blessed with a completely negative family history of either type I or type II diabetes.
So why does this matter? One of the answers has to do with what could be characterized as the utility of the information. In the MD Perspective for both type I and type II diabetes on the 23andme website both physicians acknowledge that there is currently no way to use the genomic predisposition to prevent diabetes. From the medical perspective the testing has no utility, in that the results of the test do not lead to alterations in care that improve the patient’s health outcome. This type of information is frequently used by insurers or health systems to decide whether or not a given test should be covered or even offered to patients.
Personal Utility
There is another concept that is beginning to emerge in discussion of genomic testing which is the idea of personal utility. What will this test result mean to me as a person? What will I do if the result is positive or negative? If there is no insurance coverage what am I willing to pay out of my own pocket because I believe this will help me? This principle clearly operates in medicine all the time. The supplement and nutriceutical industry not only survives but thrives as a multi-billion dollar industry almost solely on the basis of personal utility, given that there is limited evidence of increased health benefit (and despite several instances of overt harm. Fen-Phen anyone?)
Folate Flashback
I took a daily Folate supplement for a couple of years because I knew: 1) I didn’t eat as many green vegetables as I should; 2) Folate decreases serum homocysteine levels; 3) elevated homocysteine is associated with an increased risk of cardiovascular disease; 4) Even if it doesn’t help, it surely won’t hurt (and it’s easier than eating vegetables!!); 5) It’s cheap. As Folate was systematically studied, it was found that for reasons that still are unexplained, men who had higher Folate intake had more cardiovascular events. Into the trash with my Folate pills!!
Bottom line was my decision to spend discretionary income on Folate was based on a belief that Folate supplements would reduce an outcome of concern for me (a heart attack) at an acceptable ‘cost’ (price, lifestyle, added bother to remember to take the pill, etc.). For me, the ‘cost’ of actually losing weight, eating more healthily, exercising regularly, which does have demonstrated utility, is too high. My perception of the personal utility of these health behaviors is too low to justify spending time and effort to achieve them, despite abundance evidence to the contrary.
The Patient Perception Point
So what’s the point? (Aha, you assume I have one). From my perspective the issue is that we frequently ignore the disconnect between evidence of utility and a patient’s perception of utility. This leads to recommendations based on sound evidence falling on apparently deaf ears. If we make no attempt to understand what the patient is interested in and why they are interested in it, we may miss opportunities to improve health in ways we don’t anticipate. I now know that Grant is most worried about Prostate Cancer, Type II diabetes and Venous Thromboembolism. Let’s assume that Grant’s health behaviors aren’t as good as they should be. If he were overweight, would his concern about Type II diabetes based on his 23andme results present a teachable moment about what he could do to control his risk?
No way to know for sure, but it’s more likely than if his major worry was Crohns disease. I might say, Gee, Grant, you’re really concerned about diabetes and blood clots. Did you know that being overweight really increases your risk for both of those problems? In fact, losing weight has been proven to dramatically reduce your risk of developing diabetes. Is this something you’re ready take on, because if you are, I have a dietician that can help you assess your diet, and I have a list of weight loss programs that will help you to not only lose the weight, but will help you to keep it off.
The bottom line for me is listen to the patient, read between the lines, get to the patient’s concern and help them to develop a plan to address the concern. If it takes a 23andme test to get to the concern for that patient, I can live with that—and maybe they can too!
And now for something completely different
This is a complete non-sequitur, but as I was playing in Grant’s 23andme site (he lets me log in for the purposes of the blog and I only look at the topic under discussion) I found a tab called timeline. Clicking on this gives a timeline of major discoveries for the disease of interest. In the case of diabetes this goes back to 1550 BC to a citation in the Papyrus Ebers by an Egyptian physician Hesy-Ra describing sweet urine and a variety of dietary cures and ends in 2007 with the Wellcome Trust case control genome-wide association study for type I diabetes. I’m showing my geekitude (geekosity?, geekness?), but I thought it was cool.
Nice post about genetic testing for diabetes type 1 which is increase day by day. Thanks for sharing the informative knowledge. Keep me more updates.
ReplyDeleteWhat test numbers tell me if I have diabetes or prediabetes?destructeur diabète and Maitrisez Votre Diabete
ReplyDeleteI must admit ( ( Robinson.buckler (@) Yahoo. com ) ) is really a genuine man when it comes to getting an ex back
ReplyDeleteI and my friends watch the football game clips at YouTube for all time, because they have in nice quality.
ReplyDelete-------------------------
Car Transport in Pune
I can't say anything but thanks for this surprising article. It gives another thing to my mind.
ReplyDelete--------------------------
Best SEO Company In Kanpur india
The heritability of diabetes type one is too high!
ReplyDeletemoney affirmations
golden goose
ReplyDeletegolden goose shoes
adidas yeezy
yeezys
yeezys
goyard handbags
yeezy
adidas yeezy
goyard
golden goose outlet
jordan 13
ReplyDeletekyrie 8
yeezy
off white
paul george
air jordan
fear of god essentials hoodie
off white outlet
kd13
nike dunks