Commentary Article
The American Diabetes Association
(ADA) suggests tailoring medical nutrition therapy (MNT) to treat patients with
diabetes because there is no single best dietary strategy [1]. This is a reasonable
approach to gaining trust and building relationships with patients, and
possibly true given the variation among individuals. It is certainly a step
back from population-based and a step toward more holistic approaches. It would
theoretically increase compliance and overall health if some control is left up
to each individual.
The ADA mentions several times the
dangers of processed foods. One common denominator of successful eating plans
seems to be the avoidance of processed foods. Although it is unclear whether
processing the foods is harmful, if the typical ingredients in processed foods
are harmful, or if there is some combination of those two factors. Nonetheless,
authors of the SoC-25 encourage to “limit processed foods and foods high in added
salt, sugars, and fats and, when possible, choose whole foods” and to lose
weight loss of 3-7% in those who are overweight or obese [1].
The Keto-Med trial [2] was added as
a reference to the SoC-25 as evidence that a Mediterranean diet (Med-Plus) is
comparable to a well-formulated ketogenic diet (WFKD) in lowering A1c but
without the downside of increasing low-density lipoprotein (LDL)-cholesterol
(LDL-C). The implication is that by decreasing LDL-C through dietary
modification, cardiovascular disease (CVD) risk will also decrease. A
meta-analysis shows improvement in the CVD risk factors weight, triglycerides,
and high-density lipoprotein (HDL) but warns that increases in LDL-C may negate
those benefits [3]. This is the position that the ADA and the authors of the
Keto-Med trial are taking, and likely the rationale for following the ADA’s
position on diet for decades.
Nevertheless, some issues arise
from inserting this one study into the SoC-25:
·
It is unclear if
prioritizing LDL-C over all other available metrics is the best method to
assess CVD risk in the general population, diabetic population, or those
following an LCD. The issue is contentious, though there is growing data to
support LCD as a safe and effective MNT for diabetes therapy [4-6].
·
The authors of
the Keto-Med Study mention that while LDL-C increased in the WFKD arm when
compared to the Keto-Plus arm, diets lower in carbohydrates tend to result in
larger, more buoyant LDL particles and that no lipid subfractionation studies
had been performed. This was mentioned in the Keto-Med Study as a limitation
but not in the SoC-25.
·
The authors of
the Keto-Med Study acknowledge that there are other factors that improved more
on the WFKD that may contribute more to lowering CVD risk than LDL-C does. There
was a statistically significant lowering of triglycerides and CGM average
glucose and an insignificant improvement in fasting insulin, HDL, and weight in
the WFKD arm when compared to the Med-Plus arm, which was not mentioned
anywhere in the SoC-25.
·
A1c decreased in
both arms, however the difference between arms was statistically insignificant.
This contrasts with changes in CGM average glucose, which had a difference of
-8 mg/dL in the WFKD arm and -2 mg/dL in the Med-Plus arm. This mismatch between
A1c other metrics such as CGM needs to be further studied as A1c may be
artificially high in individuals that consume lower amounts of carbohydrates.
·
“We are moving
away from emphasizing macronutrients, which include carbohydrates, proteins,
and fats, and micronutrients, which include vitamins and minerals, and instead
focusing on foods.” Micronutrients are a different story. Sodium intake of
<2.3 g/day is recommended in the SoC-25. This is consistent with the DASH
diet but would be inadequate in other eating plans. Keto-Med participants in
the WFKD arm were given 3-5 g/day sodium and 3-4 g/day potassium.
Given the above improvement in more
CVD risk factors overall and the lack of need to restrict sodium intake, it
appears that the WFKD arm outperformed the Med-Plus arm. However, the bias
created by fear of LDL-C is tremendous. That idea took root after
cholestyramine for primary prevention in a 1984 NHLBI study simultaneously
lowered LDL-C and decreased CVD-related death by 0.4% (NNT=250) over 7.4 years [7].
Since then, it has been assumed that diets that lower LDL-C would provide the
same benefit.
Never mind the data from large
randomized controlled trials (RCTs) that refute the diet-heart hypothesis. The
Minnesota Coronary Survey was a huge RCT that could have been published in the
1970s but was not published until 1989 after the principal researchers retired
because “we didn’t like the way it turned out” [8-10]. The Women’s Health
Initiative showed that more harm may be done than good by lowering fat intake
and that metrics associated with insulin resistance are most predictive of CVD
risk in certain groups of women [11,12].
While the American Heart Association (AHA) has engaged in a war on saturated fat intake to lower LDL-C and hopefully lower CVD risk, the ADA has dutifully followed. There is variability between individuals that gets ignored by population medicine and may be causing more harm. A holistic approach may benefit the individual. More recent data suggest that in those starting an LCD, elevations in serum LDL-C may be related to weight rather than to saturated fat intake and that LDL-C either stays stable or decreases in obese individuals [13]. The Lipid Energy Model has been used to explain the observation [14]. LDL-C has been noted to increase after starting a VLCD even in the context of lower saturated fat intake in some individuals [15]. Increasing saturated fat and carbohydrate intake could lower LDL-C [15]. More research is needed to understand this relationship between carbohydrate restriction and lean mass individuals to the other conventional markers of CVD risk that tend to improve.
Table 1: Comparing Carbohydrate Content between the Keto-Med Study and the JLCDPA.
While the ADA appears to be
shifting toward treating the individual in a more holistic approach rather than
population-based treatment plans, there is a curious revision. Recommendation
5.29 was added to the SoC-25 to encourage increasing plant-based proteins at
the expense of animal-based proteins. The ADA admits there is little evidence
to support limiting animal protein, but that less consumption of saturated fat
might lower LDL-C and somehow save the planet. The rationale behind the latter
is not explained as the planet, presumably, is not afflicted with diabetes. The
former should be discussed with each individual. If eating cookies lowers LDL-C
[15], joint decision-making should include discussing whether or not eating
more of them is in the best interest of the patient. Likewise, lipid-lowering
medication can be prescribed depending on the whole clinical picture and the
wishes of the patient.
In a 2019 consensus report the ADA
began recognizing LCD eating patterns as a viable MNT for diabetes. They are
among the most studied eating patterns for type 2 diabetes [16]. Since then the
ADA has been consistent that MNT should be individualized. The authors agree
with this sentiment and have an interest in treating the patient rather than
the illness. We also recognize the benefits of multiple different types of
eating patterns including the Mediterranean diet [17]. Fortunately, there are
multiple methods to measure glycemic control, and it allows us to give more
control to the individual. The Japan LCD Promotion Association (JLCDPA) and
Takao Hospital recommend various degrees of LCD for this reason [18] to allow
for flexibility based on the needs of each individual. When considering the
carbohydrate composition, the Med-Plus diet (38% carbohydrate) is close to our
petit LCD (40% carbohydrate) and the WFKD (12%) matches our super LCD (12%)
[2,18]. Nuance based on the whole clinical picture as well as patient
preferences should be considered. It is exciting to see the progression of more
patient-centric thinking in healthcare.
Conflict of Interest
The authors declare no conflict of
interest.
Funding
There was no funding received for
this paper.
References