I often see people claim on social media that glucose spikes above 110/120/140 mg/dl (roughly 6/7/8 mmol/l) cause damage and diabetics should religiously keep their blood sugars below this level to prevent long term complications.
While the research to verify this assertion could well be done with the wealth of data now captured by continuous glucose monitors (CGMs), to my knowledge, it has not been done. My concern is that being this fixated on your glucose levels would be a great way to drive yourself crazy and a tragedy if it was for no benefit.
So what does science know? In this article I will review the literature as well as show you what to look for in other medical papers. As usual, feel free to go to tl;dr if reading scientific paper summaries is not your thing.
My first source is a recent study from Sweden which suggests there is a ‘goldilocks zone’ for diabetics where the HbA1c is not too high to cause complications and not too low to increase the risk of severe hypoglycemia. Their conclusion is an HbA1c between 6.5% and 6.9% is optimal to avoid these two extremes.
My second list of sources come from Blood Sugar 101. This is a site that claims their cited scientific papers “make a cogent case that post-meal blood sugars of 140 mg/dl … cause both permanent organ damage and the worsening of diabetes.” I am keen to review their papers to see if the papers actually support this position. I have nothing against the site, I barely know it. I chose it simply because it was cited in social media and I assume they have chosen papers to give the most compelling case for their claim.
What To Look For In Medical Papers
When reviewing papers and their findings, I look at two things: their ‘n’ and ‘p’ values. The ‘n’ is the number of people involved in the study (obviously the bigger, the better) and the ‘p’ value which measures statistical significance. The lower the ‘p’ value, the more reliable the conclusions with a value below 0.05 generally considered to be statistically significant.
For example, the Swedish study mentioned above had n=10,398. That is quite a big study. The Results section says the following:
“Mean age of participants was 14.7 years (43.4% female), mean duration of diabetes was 1.3 years, and mean HbA1c level was 8.0% (63.4 mmol/mol). After adjustment for age, sex, duration of diabetes, blood pressure, blood lipid levels, body mass index, and smoking, the odds ratio for mean HbA1c <6.5% (<48 mmol/mol) compared with 6.5-6.9% (48-52 mmol/mol) for any retinopathy (simplex or worse) was 0.77 (95% confidence interval 0.56 to 1.05, P=0.10), for preproliferative diabetic retinopathy or worse was 3.29 (0.99 to 10.96, P=0.05), for proliferative diabetic retinopathy was 2.48 (0.71 to 8.62, P=0.15), for microalbuminuria or worse was 0.98 (0.60 to 1.61, P=0.95), and for macroalbuminuria was 2.47 (0.69 to 8.87, P=0.17). Compared with HbA1c levels 6.5-6.9%, HbA1c levels 7.0-7.4% (53-57 mmol/mol) were associated with an increased risk of any retinopathy (1.31, 1.05 to 1.64, P=0.02) and microalbuminuria (1.55, 1.03 to 2.32, P=0.03). The risk for proliferative retinopathy (5.98, 2.10 to 17.06, P<0.001) and macroalbuminuria (3.43, 1.14 to 10.26, P=0.03) increased at HbA1c levels >8.6% (>70 mmol/mol). The risk for severe hypoglycaemia was increased at mean HbA1c <6.5% compared with 6.5-6.9% (relative risk 1.34, 95% confidence interval 1.09 to 1.64, P=0.005). “
It looks complicated but we can break it down. When it compares the risk of any retinopathy between those with an HbA1c < 6.5% and those with 6.5-6.9% the odds ratio is 0.77 (you are less likely to get retinopathy with the higher HbA1c) BUT the ‘p’ value is 0.10 so it is not statistically significant and we can ignore it. In fact, in comparing the <6.5% group to the 6.5-6.9% group, the only statistically significant result was for preproliferative diabetic retinopathy which was right at the edge of significance (p=0.05).
However, when comparing <6.5% to 7.0-7.4% and >8.6%, across the board, there was a statistically significant increase in risk for all of the examined complications.
Finally, when comparing the risk of severe hypoglycemia between <6.5% and 6.5-6.9% there was a statistically significant increase in risk below 6.5% (34% higher).
The paper’s conclusion is:
“Risk of retinopathy and nephropathy did not differ at HbA1c levels <6.5% but increased for severe hypoglycaemia compared with HbA1c levels 6.5-6.9%. The risk for severe complications mainly occurred at HbA1c levels >8.6%, but for milder complications was increased at HbA1c levels >7.0%”.
This makes sense and I believe the “mainly” is inserted to cover the borderline preproliferative diabetic retinopathy risk increase for the 6.5-6.9% group.
Now let us look at the case for “140 mg/dl does damage” by going through the Blood Sugar 101 sources.
Reviewing the Papers
Some of the Blood Sugar 101 links were broken but here are the ones which actually went somewhere or which I could find by Googling the title.
n=107 of which only 13 had diabetes and 36 had impaired glucose tolerance (IGT) and all had idiopathic (unknown cause) neuropathy.
The paper found people with IGT (defined as having a blood glucose of 140-200 after two hours in an oral glucose tolerance test (OGTT)) had a statistically significant higher change of having neuropathy BUT no such conclusion was made for diabetics. In other words, the low population of the study, combined with the low population of diabetics means this paper offers little value to diabetics and yet I have seen it quoted on a few sites claiming it backs the “over 140 mg/dl does damage” claim. At best, we can say it supports the claim that people who are prediabetic are at a greater risk of neuropathy, but that is about it.
This study mirrored the previous one with n=73. Of these patients, 26 had IGT, and 15 had diabetes. This paper shows that diabetics that have neuropathy have it more severely than those just with IGT. So, in this case, the conclusion is if someone is at 200 after two hours of an OGTT (the definition of ‘frank’ diabetes) if they get neuropathy it will likely be more severe than their prediabetic counterparts.
In this one n=100, all with chronic idiopathic axonal polyneuropathy (CIAP). They were given an OGTT and 62 of them had abnormal results, twice as high as general population groups. Statistical significance was a little light on the ground in this study but it is aligned to the previous two studies’ findings.
This study had n=195 diabetics and n=198 control subjects. It found diabetes was a risk factor for polyneuropathy and, within the diabetic group, age, waist circumference, and peripheral arterial disease were associated with polyneuropathy.
This study tried to keep n=800 critically ill patients patients below 140 mg/dl while in the Intensive Care Unit (ICU) over a period of 11 months and compared them to patients who were not intensively managed. The populations were not all diabetic with the only common factor being admission to ICU.
The following were shown to have a decreased incident rate in the intensively managed patients: poor kidney function (renal insufficiency), blood transfusions, hospital mortality rates, and length of stay in the ICU. Hypoglcemia rates did not significantly change.
This one is in mmol/l but I will convert for the US diabetics. Essentially it showed that when blood glucose goes above 100 mg/dl, the ratio of insulin sensitivity to insulin resistance declined. However, the paper failed to report the level of statistical significance of the results. It did say it used n=388 though of which 250 had IGT or Type 2 diabetes. So, assuming the results were significant, it tells us that either resistance increases or sensitivity decreased as blood glucose levels go up.
This study reviewed the beta cell mass of bodies from 124 autopsies. Of the 124, 91 were obese and 33 were lean. They found that obese patients had roughly a 50% larger beta cell volume (possibly influenced by the younger age at which the obese population died). Of the obese individuals, the Type 2s had a 63% smaller beta cell volume than their non-diabetic obese counterparts.
The rest of the paper talks at the possible mechanisms for this difference is volumes, looking at beta cell replication rates and beta cell death rates.
This is a mice study and, given the number of ‘cures’ for diabetes found for mice, I am a little skeptical to apply the findings to humans. The paper was looking at the survival rate of transplants between mice with insulin treatment to keep their glucose below 150 mg/dl and mice with no such treatment.
The paper found:
“…insulin treatment did not improve the initial preservation of transplanted β-cell mass in the initial days after transplantation. In contrast, increased apoptosis (cell death) and reduced β-cell mass were found in islets exposed to long-term hyperglycemia but not in normoglycemic mice, suggesting that sustained hyperglycemia increased β-cell death in transplanted islets.”
So transplanted beta cells in mice did not appreciate long term exposure to elevated glucose levels.
This was a study of cells growing in a culture at different glucose levels with no link made back to human glucose levels.
This study gave n=1062 patients an OGTT and measured their blood glucose after one hour. Those above 155 mg/dl had elevated inflammatory markers and lipid ratios. The author goes on to suggest these increases could be a marker for cardiovascular risk but does not provide evidence linking the markers to heart disease.
This was a link to another Blood Sugar 101 page which had a bunch more links but, given the length of this article already, I am focusing on the ones just on the original page. If enough people call this out, I am happy to review the heart disease one in another article.
Broken link and could not find the source on Google. I did find this summary but without indication of statistical significance it is hard to confirm the findings. Also, the paper focused on pre-diabetes so its relevance to diabetics is limited especially when no blood glucose levels are mentioned. I expect it found conclusions similar to papers (1), (2), (3), and (4).
This paper looked at the data of three populations (n=3162, n=2182, and n=6079). Its conclusion was:
“We saw no evidence of a clear and consistent glycaemic threshold for the presence or incidence of retinopathy across different populations. The current FPG cutoff of 7·0 mmol/l used to diagnose diabetes did not accurately identify people with and without retinopathy.”
In other words, they found that a person’s fasting plasma glucose (FPG) was a poor predictor of retinopathy.
This paper is looking at FPG and HbA1c to see if it is predictive for diabetic retinopathy. With an n=1066 (not all diabetics) they concluded that the greatest increase in prevalence for retinopathy occurred for HbA1c above 5.5% and FPG above 5.8 mmol/L (105mg/dl). It also found that HbA1c was a better predictor than FPG. Here are their curves.
For the HbA1c curve, while the uptick is at 5.5%, we see the dip before this means the prevalence, relative to the baseline prevalence of around 10% only starts inceasing past this above 6%. Similarly, to escape baseline required an FPG above around 6.5 mmol/l
This study combined the results of nine studies to get a whopping n=44,623. They looked at FPG (n=41,411), two-hour OGTT (n=21,344), and HbA1c (n=28,010).
While no ‘p’ values were given, their results concluded that an HbA1c above 6% has an increased prevalence of retinopathy with the threshold for significant risk at above 6.4%. For FPG the threshold was 6.6mmol/l (120 mg/dl). OGTT proved to be a poor predictor.
n=700 with the aim to determine the HbA1c and FPG for predicting retinopathy after 10 years.
Here are the results.
While the paper’s conclusions were thresholds of 108 mg/dL for FPG and from 6.0% for HbA1c, we can see above the prevalence only jumps up significantly after >7.0 mmol/L (126 mg/dl) for FPG and >7.0% for HbA1c.
This was a press release talking about two studies, rather than the studies themselves. Given it is light on details, I am ignoring it for this analysis. It did say this though:
“No one is claiming, based on current evidence, that either fasting glucose or HbA1C is a viable target for therapy of heart failure specifically; that would have to be established in prospective, randomized trials, all three researchers emphasized.”
This paper looked at n=33,293 women and 31,304 men (for a total of n=64,597). Of these, 2,478 people had cancer. The big takeaway of this study was the difference in risk profile between men and women. It found “abnormal glucose metabolism was associated with a statistically significantly increased risk of cancer overall in women but not in men.”
To put it another way: “In men, overall, no statistically significant associations were observed between glucose levels and cancer risk”.
Like study (9), this was a study of cells in a lab, rather than a study of humans. The main conclusion was a fluctuation in glucose levels aligned to the kinds of fluctuations a human body is exposed to through three meals a day and 12 hours of fasting is more damaging than constantly high glucose levels.
n=1871, all diabetics, had their HbA1c measured and then were followed up over a period of 11 years. The groups were split into people with HbA1cs of <6%, 6-7%, 7-8%, and >8%. Groups above <6% had a higher relative risk of chronic kidney disease (CKD).
This means, if there is a threshold for HbA1c, above which CKD begins to increase in risk it probably lies somewhere between 6 and 7%.
This study involves n=19,019 men but the analysis only looked at the non-diabetic ones (n=18,406). The men did a test similar to an OGTT but not quite following modern protocols and if they exceeded 200 mg/dl they were excluded (n=56). Also those with missing data were excluded (n=134) leaving a total of n=18,216.
The study found that for non-diabetics, the risk of stroke mortality increased if the patient’s blood sugar went over 4.6mmol/l (82.8 mg/dl) as part of their pseudo-OGTT. Given my focus is on diabetics, a paper studying non-diabetics is of limited relevance.
Summary Of All The Paper’s Conclusions
So this is what we know from the 21 papers.
- If you have impaired glucose tolerance (IGT) you are more likely to get neuropathy
- Diabetics who fail an OGTT are at risk of more severe neuropathy than those with just IGT
- If you have IGT you are more like to get chronic idiopathic axonal polyneuropathy (CIAP)
- Diabetes is a risk factor for polyneuropathy
- If you use insulin to keep patients in intensive care under 140 mg/dl, they tend to fare better
- Either insulin resistance increases or sensitivity decreases as blood glucose levels go up (this is anecdotally confirmed by Type 1s I know who say it takes much more insulin to come down from a large high than a smaller one.)
- Obese Type 2s have a smaller beta cell volume than their obese non-diabetic counterparts
- Keeping glucose levels lower in mice with pancreatic transplants improves the longer term prospects of the transplant
- Cells in a dish do not like higher glucose levels
- People with IGT get inflammation when they spike
- Study not found but likely found that people with IGT are at a higher risk of neuropathy
- Fasting Plasma Glucose (FPG) is a poor predictor of retinopathy
- HbA1c is a better predictor of retinopathy than FPG and to get above the baseline risk, required an HbA1c of greater than 6% or a FPG of 6.5 mmol/l (around 120 mg/dl)
- HbA1c above 6% has an increased prevalence of retinopathy with the threshold for significant risk at above 6.4%. For FPG the threshold was 6.6mmol/l (120mg/dl). OGTT proved to be a poor predictor.
- The risk for retinopathy significantly increases when HbA1c is above 7% or when FPG is above 7 mmol/l (around 130 mg/dl). There is a smaller increase when the HbA1c is above 6.5%
- Neither HbA1c nor FPG were seen as viable targets for heart failure therapy
- Women with high glucose levels are at a greater risk of getting cancer
- Cells in a lab will tolerate high levels of glucose exposure better than fluctuating levels of glucose
- If there is an HbA1c threshold for chronic kidney disease, it is probably somewhere between 6-7%
- A study exclusively focusing on non-diabetics using a non-standard OGTT. Therefore it is of limited relevance to Type 1s
You will see above that not one of these papers directly examined blood spikes over 110/120/140 mg/dl. The 140 mg/dl probably comes from the OGTT where IGT is defined as someone who has a blood glucose of 140 mg/dl after two hours. This says nothing about the number of glucose spikes a patient has had before the test or how high those spikes went. The 140 mg/dl limit in an OGTT tells us nothing about the level at which individual glucose spikes do damage to the body.
The 120 mg/dl may come from (15) where it was shown that an FPG above this led to an increased risk of retinopathy but a fasting glucose level says nothing about someone who is below this FPG level and occasionally spikes above 140 mg/dl.
As for the 110 mg/dl, I have no idea where this one comes from. Regardless, none of the 21 references provided evidence to support a “cogent case” that occasional spiking leads to long term damage.
So What Can We Conclude?
Summarizing my summary and including my original Swedish study, we get the following in regards to Type 1s and what blood levels make sense to stay healthy:
- (Swedish study) An HbA1c below 6.5% increases the risk of severe hypoglcemia
- (Swedish study) An HbA1c above 6.9% increased the risk of complications, including retinopathy
- (Swedish study) There is an increased risk of preproliferative diabetic retinopathy above an HbA1c of 6.5%
- (1), (2), (3), (4), (12) People who fail an OGTT have an increased risk of neuropathy
- (13), (14), (15) FPG is a poor predictor of retinopathy but risk appears to increase above 120 mg/dl
- (14), (15) HbA1c is a better predictor of retinopathy and risk increases above 6%, with significant risk above 6.4%
- (16) There is a small increase in risk of retinopathy for an HbA1c above 6.4% with a significant risk above 7.0%
- (16) An FPG above 130 mg/dl increases the risk of retinopathy
- (20) If there is an HbA1c threshold for chronic kidney disease, it is probably somewhere between 6-7%. If there is no threshold, an HbA1c above 6% increases the risk
Clearly Fasting Plasma Glucose (FPG) and HbA1c have multiple studies examining at what point a diabetic has an increased risk of complications with retinopathy being a common complication studied.
Based on the above, it is clear that an HbA1c below 7.0% is desirable (Swedish study, (16)) and, likely, an HbA1c below 6.4% is better (Swedish study, (14), (15), (16), (20)). However, an Hba1c below 6.4% does put an insulin-dependent diabetic at an increased risk of a severe hypo (Swedish study) so, therefore, depending on how well you can manage the fluctuations may determine where your target HbA1c range will sit.
For Fasting Plasma Glucose, while not as strong a predictor as HbA1c, keeping it below 120 mg/dl would be prudent (13), (14), (15), (16).
Reviewing the multiple studies of a site which makes the claim that blood sugars above 140 mg/dl cause damage and worsen diabetes, not one directly studied meal spikes and their long term effects.
However, when these studies were combined with a recent Swedish study, we can conclude that keeping your HbA1c below 7.0%, and for those who have a low risk of hypo, below 6.4% will minimize the risk of complications, especially retinopathy.
For fasting plasma glucose (FPG), keeping this below 120 mg/dl (6.7 mmol/l) is also desirable to reduce the risk of complications but it should be acknowledged that FPG is not as reliable as a predictor of complications as HbA1c.
Finally, given there was no study examining the damage of meal spikes and assessing a ‘safe’ level, it is reasonable to ask whether religiously guarding your blood glucose levels is worth it; whether the mental fatigue of constant monitoring, and risk of burnout, is outweighed by the unproven benefits. Perhaps it is better to focus on longer term measures such as HbA1c, the standard deviation of glucose levels over time, and time in range. Perhaps it is better to see an occasional high spike as an unfortunate day on a much longer journey rather than as a defeat or failure.