Very low-carbohydrate diets in the management of diabetes revisited - Semantic Scholar
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
VIEWPOINT
Very low-carbohydrate
diets in the management of
diabetes revisited
Grant Schofield, George Henderson, Simon Thornley, Catherine Crofts
ABSTRACT
Humans can derive energy from carbohydrate, fat, or protein. The metabolism of carbohydrate requires by
far the highest secretion of insulin. The central pathology of diabetes is the inability to maintain euglycaemia
because of a deficiency in either the action or secretion of insulin. That is, because of either insulin resistance
often accompanied by hyperinsulinaemia, or insulin deficiency caused by pancreatic beta cell failure. In
individuals dependent on insulin and other hypoglycaemic medication, the difficulty of matching higher
intakes of carbohydrates with the higher doses of medication required to maintain euglycaemia increases
the risk of adverse events, including potentially fatal hypoglycaemic episodes. Thus, mechanistically it has
always made sense to restrict carbohydrate (defined as sugar and starch, but not soluble and insoluble
fibre) in the diets of people with diabetes. Randomised clinical trials have confirmed that this action based
on first principles is effective. The continued recommendation of higher-carbohydrate, fat-restricted diets
has been criticised by some scientists, practitioners and patients. Such protocols when compared with
very low-carbohydrate diets provide inferior glycaemic control, and their introduction and subsequent
increase in carbohydrate allowances has never been based on strong evidence. The trend towards higher-
carbohydrate diets for people with diabetes may have played a part in the modern characterisation of type
2 diabetes as a chronic condition with a progressive requirement for multiple medications. Here we will
introduce some of the evidence for very low-carbohydrate diets in diabetes management and discuss some
of the common objections to their use.
D
ata from the 2008/2009 New Zealand to meet saturated fat recommendations.3
Nutrition Survey placed the prev- Although weight control is important for
alence of diabetes among adults at diabetes prevention, the ‘Topical Questions
7%. The incidence of pre-diabetes, defined and Answers’ supplement to the guidelines
as HbA1c, 41 to 49 mmol/mol (5.8% to is briefly dismissive of low-carbohydrateVIEWPOINT
diabetes. We also present some evidence a focus on avoiding specific foods instead.
supportive of benefit from carbohydrate For example, in the very low-carbohydrate
restriction in type 1 diabetes. pilot trial of Unwin et al (2015), patients
We accept that there are useful methods with type 2 diabetes (n=68) were advised
of managing diabetes aligned with carbo- to avoid sugar, bread, pasta, rice, and other
hydrate restriction, including glycaemic carbohydrate-dense foods.9
index (GI), Mediterranean diet, ancestral
diets designed on evolutionary principles, Is dietary
and drugs such as metformin and acarbose.
However, our focus will be on carbohy-
carbohydrate
drate restriction itself and the relative necessary?
effect of carbohydrate and fat as sources of Some authorities consider dietary carbo-
energy. The basic principle of carbohydrate hydrate essential to provide fuel, especially
restriction in the management of diabetes for the brain, and are concerned about
underlies many of these other approaches, an increased possibility of hypoglycaemia
and sufficient clinical evidence supports its or ketoacidosis when carbohydrate is
use. We highlight evidence relating to very restricted. Although very low-carbohy-
low-carbohydrate diets, to best illustrate the drate diets still supply some carbohydrate
mechanisms and effects of carbohydrate from non-starchy vegetables and low-sugar
restriction. fruits, as well as optimal amounts of fibre,
there is little evidence that dietary carbo-
What is a very hydrate is essential in the human diet.10
low-carbohydrate The claim sometimes made that carbohy-
drates are necessary in the diet of people
diet? with type 2 diabetes is also not supported
Studies differ in the exact degree of by the physiology of the disease. Type 2
carbohydrate restriction tested for effect diabetes is characterised by a higher than
on glycaemic control. Feinman et al (2015), usual hepatic production of glucose from
in their critical review of carbohydrate amino acids and glycerol.11 In normal
restriction for diabetes, proposed that “very metabolism this production of glucose
low-carbohydrate” be defined as 20–50g/ is reduced whenever carbohydrate is
day of carbohydrate.5 This is an attainable consumed, but in type 2 diabetes it is not,
figure rooted in the pre-insulin history and this, as well as delays in peripheral
of diabetes and the physiology of carbo- glucose clearance due to insulin resistance,
hydrate metabolism.6,7 Setting the goal results in post-prandial and fasting serum
carbohydrate intake as low as practicably glucose elevation.12 A person with type 2
possible is consistent with the meta-analysis diabetes who is not using glucose-lowering
of randomised controlled trials of low- medication is thus an unlikely candidate
carbohydrate diets for diabetes by Kirk for symptomatic hypoglycaemia on a low-
et al (2008), which found that levels of carbohydrate diet. If glucose-lowering
fasting serum glucose, HbA1c, and trig- drugs are being used, then doses will need
lycerides were inversely proportional to to be adjusted appropriately (an all-im-
the percentage of carbohydrate consumed portant caveat).5 Glucose requirements are
in each diet phase.8 Diets that exceed the also reduced by the switch to utilising fat
putative normal carbohydrate requirement and ketone bodies for fuel when carbohy-
of 130g/day should not be considered drate is unavailable for prolonged periods.7
‘low-carbohydrate', as the desired physio- Ketone bodies are sufficiently supportive of
logical changes (lowered insulin demand brain function and neuronal health that it
or secretion, decreased appetite, inhibition is well-recognised that a ketogenic diet is a
of lipid synthesis, and increased fatty acid viable treatment option for certain forms of
oxidation) depend on adaptations to carbo- epilepsy and other neurological disorders.13
hydrate deprivation.7 Type 1 diabetes results from the auto-
An absolute or percentage of energy immune destruction of beta cells, which
carbohydrate limit can sometimes be no longer supply insulin to lower serum
omitted altogether in clinical practice, with glucose and inhibit lipolysis. A diagnosis
68
NZMJ 1 April 2016, Vol 129 No 1432
ISSN 1175-8716 © NZMA
www.nzma.org.nz/journalVIEWPOINT of type 1 diabetes is often a medical emer- composition to today’s diets.6 The efficacy gency and will require the life-long use of this diet was shown in clinical practice of exogenous insulin, whereas people and its safety in terms of blood sugar, blood with type 2 diabetes are potentially able lipids, and ketoacidosis was tested experi- to reverse dependence on medication mentally throughout the 1920s and 1930s. through diet and lifestyle changes. Insulin The first high-carbohydrate diet for diabetes is the hormone that facilitates the removal that did not carry a high risk of diabetic of glucose from the bloodstream, and the ketoacidosis dates from 1926 and required use of insulin introduces a risk of hypo- the initiation of insulin treatment.16 Low- glycaemia if insulin dose is not matched carbohydrate, high-fat diets therefore have exactly to food intake. People with insu- a long history of usefulness in the treatment lin-dependent diabetes, including those on of diabetes and only went out of fashion very low-carbohydrate diets, are advised to when the diet-heart hypothesis (proposing use glucose tablets or ‘emergency foods’ to a link between saturated fat and heart treat hypoglycaemic episodes.14 disease) emerged. The safety net provided Although rare cases of ketoacidosis have by glucose-lowering drugs, earlier diagnosis been reported in people without diabetes of type 2 diabetes, and increasing access using very low-carbohydrate diets for to self-monitoring equipment provided a weight loss, we have only found one report context that allowed higher-carbohydrate of ketoacidosis in a person with diabetes diets to flourish. To our knowledge, no trials using a low-carbohydrate diet. A 32-year-old that compared very low-carbohydrate and woman with Prader-Willi syndrome and low-fat, low-saturated fat (
VIEWPOINT
of glycaemia with insulin in people with low-carbohydrate diet with the conven-
type 1 diabetes. In one small case-series tional low-fat diet for type 2 diabetes over
of intensive glycaemic control with drugs 52 weeks.20 In this trial, in which satu-
and diet, subjects diagnosed with type 1 rated fat was restricted, lipids, including
diabetes (n=10) and type 2 diabetes (n=20) low-density lipoprotein cholesterol (LDL-C)
were placed on the same very low-carbo- improved on the very low-carbohydrate
hydrate diet (30g/day carbohydrate).14 diet. In a 2008 very low-carbohydrate
HbA1c in those with type 1 diabetes was weight-loss randomised controlled trial
reduced in all cases, the mean reduction in obese subjects without type 2 diabetes
being from 6.8% to 5.5% (50.8 to 36.6 mmol/ (n=88) by the same group, saturated fat was
mol). In participants with type 2 diabetes, not restricted (saturated fat was 20% of
the mean HbA1c reduction was from 8.4% energy) in the very low-carbohydrate arm.
to 5.8% (68.3 to 39.9 mmol/mol). No cases of High-density lipoprotein cholesterol
severe hypoglycaemia were reported over (HDL-C) and triglycerides improved, while
the study period of 21.4 months (standard LDL-C decreased overall but increased by
deviation (SD): 22.3). In a small clinical 10% in 24% of the participants.21 However,
audit (n=48) of a low-carbohydrate diet apolipoprotein beta (ApoB) did not increase
(defined as 75/g day carbohydrate) for type in this subgroup. It is notable that LDL-C is
1 diabetes, mean HbA1c in those adherent a calculated proxy for ApoB, small, dense
to the diet for the full 4-year period (48%) LDL-C particles (sdLDL), and other athero-
was reduced from 7.7% to 6.4% (60.7 to 46.4 genic factors within the LDL-C fraction.22
mmol/mol).18 The mean rate of symptomatic Evidence for the effect on cardiovascular
hypoglycaemia was reduced from 2.9 (SD: health of low-carbohydrate diets to date
2.0) to 0.5 (SD: 0.5) episodes per week. seems to show overall improvement in
glycaemic and lipid metabolic risk factor
What is the evidence surrogate endpoints compared to other
for the safety of the diets in low-carbohydrate diet trials, in only
some of which saturated fat was restricted.2
high fat content We have argued previously that analyses of
of the very low- long-term cardiovascular disease risk diet
trials which restricted saturated fat (but did
carbohydrate diet? not include low-carbohydrate arms) have
A frequently heard criticism of very not shown any major detrimental or bene-
low-carbohydrate diets for diabetes is that ficial effect of saturated fat restriction on
“eating more protein and fat may increase overall mortality.23
your risk of heart disease in the long term”. Some studies show better metabolic
It is well-known that a diagnosis of diabetes health in people who consume higher
greatly increases the risk of cardiovas- intakes of full-fat dairy foods. In the most
cular disease. It is less well-known that recent meta-analysis of epidemiological
there is also a reciprocal relationship; a prospective cohort studies of saturated
cardiovascular event predicts a future fat and trans-fat consumption, highest
diabetes diagnosis.19 We would submit that versus lowest consumption of the natu-
few things can be worse for the coronary rally-occurring trans-fats found in dairy
arteries than the metabolic and hormonal was associated with a significant reduction
changes that result in a diagnosis of type (relative risk 0.58, 95%, CI: 0.46–0.74) in the
2 diabetes. Reversing the progression incidence of type 2 diabetes.24 In a recent
of diabetes and the related metabolic prospective cohort study [n=26,930] from
syndrome should be the first priority. Malmö, Sweden, which used a 7-day food
Support for the safety of high-fat diets diary and a 1 hour interview, as well as
in people with diabetes comes from trials a food frequency questionnaire to assess
which show improved markers of cardio- dietary intake, both dairy fat consumption
vascular risk in the low-carbohydrate, (including butter and cream) and intake of
compared to the usual care group. A recent, the short- and medium-chain saturated fats
well-designed, randomised controlled trial (C4:0–C14:0) found in dairy were associated
(n=115) by Tay et al (2015) compared a very with a significantly reduced incidence of
70
NZMJ 1 April 2016, Vol 129 No 1432
ISSN 1175-8716 © NZMA
www.nzma.org.nz/journalVIEWPOINT
type 2 diabetes over 14 years of follow-up.25 signalling.28 The development of non-alco-
Lean red meat consumption was asso- holic fatty liver disease (NAFLD) leads to a
ciated with type 2 diabetes, but fatty red higher output of hepatic glucose and triglyc-
meat consumption was not. This evidence erides and is associated with an increased
seems to contradict the low fat, saturated risk of type 2 diabetes and cardiovascular
fat-restricted diabetes prevention recom- disease.29 The reduced insulin and blood
mendations of New Zealand authorities, sugar response to a very low-carbohydrate
who continue to advise the consumption of diet rapidly leads to the mobilisation and
low-fat dairy foods.3,4 oxidation of accumulated hepatic lipids.30 In
the pilot trial of Unwin et al (2015) of a very
Does the very low-carbohydrate diet for type 2 diabetes,
low-carbohydrate serum gamma-glutamyl transferase (GGT),
a surrogate marker of NAFLD, decreased
diet have any from 76.9 iu/L (95% CI: 58.3, 95.6) to 49.8
unique advantages iu/L (95% CI: 33.0, 50.3) in the patients for
whom data was available (n=64/68).9 It is
over other dietary notable that this change was not correlated
approaches? with weight loss. If metabolic improvement
is less dependent on weight loss on a
The conventional approach to diet, weight very low-carbohydrate diet compared to
loss and glycaemic control in type 2 diabetes other diets, this is a potentially important
can fairly be summarised as “all diets advantage for patients who do not lose
improve glycaemic control if they cause
weight easily, or for normal-weight indi-
weight loss, all diets that reduce energy
viduals diagnosed with type 2 diabetes.
intake cause weight loss if adhered to,
therefore there is little reason to prefer one It should be noted that low-carbohydrate
type of diet over another”. However, some diets as advised in practice necessitate
trial evidence contradicts these assertions. higher quality, micronutrient-dense food,
In a 2014 subgroup analysis of subjects with such as non-starchy vegetables, nuts and
type 2 diabetes (n=46) within a 48-week, seeds, and low-sugar fruits. The importance
well-designed, randomised controlled of micronutrient status in the management
weight-loss trial, serum glucose improved of diabetes and metabolic syndrome, and
significantly more on a very low-carbohy- the benefit from a higher intake of nutri-
drate diet compared to a low-fat diet plus ent-dense foods, are well-documented.31
orlistat, even when weight loss was equal.26 Motivated people with diabetes can
Furthermore, in a series of well-designed be advised to restrict their carbohydrate
carbohydrate restriction experiments, intake. Adherence to this advice has the
improvements in glycaemic control and potential to improve their quality of life,
hormonal and lipid parameters were demon- decrease their dependence on drugs, and
strated under conditions where patients reduce their risk of mortality based on their
were maintained at a constant weight.5,27 metabolic profile. In people with insulin-de-
One suggested explanation for these pendent diabetes, the risk of hypoglycaemic
results is that triglycerides stored in episodes is also likely to improve. As we
different parts of the body have different have outlined, this alternative approach,
effects on insulin sensitivity. In the which challenges high-starch, low-fat
aetiology of type 2 diabetes, hyperinsuli- guidelines, is supported by many lines of
naemia and hyperglycaemia result from evidence. We suggest that clinical dietary
the accumulation of triglycerides in the advice for the treatment of diabetes, as well
liver and subsequently in the pancreas, as population prevention guidelines, be
impairing endocrine and paracrine insulin urgently revised.
71
NZMJ 1 April 2016, Vol 129 No 1432
ISSN 1175-8716 © NZMA
www.nzma.org.nz/journalVIEWPOINT
Competing interests:
Nil
Author information:
Grant Martin Schofield, The Human Potential Centre, Auckland University of Technology,
Auckland; George Henderson, The Human Potential Centre, Auckland University of Technol-
ogy, Auckland; Simon James Thornley, Population Health, Counties Manukau District Health
Board, Auckland; Catherine Crofts, The Human Potential Centre, Auckland University of
Technology, Auckland.
Corresponding author:
George Henderson, The Human Potential Centre, Auckland University of Technology,
Auckland.
puddleg@gmail.com
URL:
www.nzma.org.nz/journal/read-the-journal/all-issues/2010-2019/2016/vol-129-no-1432-1-
april-2016/6850
REFERENCES:
1. Coppell KJ, Mann JI, Further observations on 2 Diabetes: Implications
Williams SM, Jo E, et al. the use of a high fat diet in for Therapy. Diabetes.
Prevalence of diagnosed the treatment of diabetes 2010;59(11):2697–2707.
and undiagnosed diabetes mellitus. Arch Intern 13. Barañano KW, Hartman
and prediabetes in New Med. 1923;31(4):455-490. AL. The ketogenic diet:
Zealand: findings from 7. Westman EC, Feinman uses in epilepsy and other
the 2008/09 Adult Nutri- RD, Mavropoulos JC, et neurologic illnesses. Curr
tion Survey. N Z Med J. al. Low-carbohydrate Treat Options Neurol.
2013;126(1370):23-42. nutrition and metabo- 2008;10(6):410-419.
2. Ajala O, English P, Pinkney lism. Am J Clin Nutr. 14. O’Neill DF, Westman EC,
J. Systematic review and 2007;86(2):276-84. Bernstein RK. The effects
meta-analysis of different 8. Kirk JK, Graves DE, Craven of a low-carbohydrate
dietary approaches to TE, et al. Restricted-carbo- regimen on glycemic
the management of type hydrate diets in patients control and serum lipids
2 diabetes. Am J Clin with type 2 diabetes: a in diabetes mellitus.
Nutr. 2013;97(3):505-16. meta-analysis. J Am Diet Metab Syndr Relat
3. Ministry of Health. Eating Assoc. 2008;108:91- 100. Disord. 2003;1(4):291-8.
and Activity Guidelines 9. Unwin DJ, Cuthbertson 15. Hayami T, Kato Y, Kamiya
for New Zealand Adults. DJ, Feinman R, Sprung VS. H. et al. Case of ketoacido-
Wellington: Ministry of A pilot study to explore sis by a sodium- glucose
Health; 2015. http://www. the role of a low-carbo- cotransporter 2 inhibitor
health.govt.nz/publication/ hydrate intervention in a diabetic patient
eating-and-activity-guide- to improve GGT levels with a low-carbohydrate
lines-new- zealand-adults and HbA1c. Diabesity in diet. J Diabetes Investig.
4. Diabetes New Zealand. Practice. 2015;4:102–8. 2015;6(5):587-90.
FAQ. [accessed 2015 10. Westman EC. Is 16. Sansum WD, Blatherwick
15 November] Avail- dietary carbohydrate NR, Bowden R. The use
able from: http:// essential for human of high-carbohydrate
www.diabetes.org.nz/ nutrition? Am J Clin Nutr. diets in the treatment of
food_and_nutrition/ques- 2002;75(5):951-953. diabetes mellitus. JAMA.
tions_and_answers#faq-8 11. Magnusso, I, Rothman DL, 1926;86(3):178-181.
5. Feinman RD, Pogozelski Katz LD, et al. Increased 17. Turner RC, Cull CA, Frighi
WK, Astrup A, et al. rate of gluconeogenesis in V, Holman RR. Glycemic
Dietary carbohydrate type II diabetes mellitus. A control with diet, sulfo-
restriction as the first 13C nuclear magnetic reso- nylurea, metformin,
approach in diabetes nance study. J Clin Invest. or insulin in patients
management: critical 1992;90(4):1323–1327. with type 2 diabetes
review and evidence base. 12. Rizza RA. Pathogenesis of mellitus. Progressive
Nutrition. 2015;31(1):1-13. Fasting and Postprandial requirement for multiple
6. Newburgh LH, Marsh PL. Hyperglycemia in Type therapies (UKPDS 49).
72
NZMJ 1 April 2016, Vol 129 No 1432
ISSN 1175-8716 © NZMA
www.nzma.org.nz/journalVIEWPOINT
JAMA. 1999;281(21):2005- Polyunsaturated fats 2 diabetes. Diabetes Obes
2012. versus carbohydrates Metab. 2014;16(1):90-93.
18. Nielsen JV, Gando C, for cardiovascular 27. Gannon MC, Nuttall FQ.
Joensson E, Paulsson disease prevention and Control of blood glucose
C. Low carbohydrate treatment. Annu Rev in type 2 diabetes without
diet in type 1 diabetes, Nutr. 2015;35:517–43. weight loss by modifica-
long-term improvement 23. Thornley S, Henderson G, tion of diet composition.
and adherence: a clinical Schofield G. Chewing the Nutr Metab. 2006;3:16.
audit. Diabetol Metab saturated fat: should we 28. Taylor R. Type 2
Syndr. 2012;4:23. or shouldn’t we? N Z Med Diabetes. Etiology and
19. Hu FB, Stampfer MJ, J. 2014;127(1394):94-6. reversibility. Diabetes
Haffner SM, et al. Elevated 24. de Souza RJ, Mente A, Care. 2013;36(4):1047-1055.
risk of cardiovascular Maroleanu A, Cozma, et 29. Yki-Järvinen H. Non-alco-
disease prior to clinical al. Intake of saturated holic fatty liver disease as
diagnosis of type 2 diabe- and trans unsaturated a cause and a consequence
tes. Diabetes Care. 2002 fatty acids and risk of all of metabolic syndrome.
July;25(7):1129-1134. cause mortality, cardio- Lancet Diabetes Endo-
20. Tay J, Luscombe-Marsh vascular disease, and type crinol. 2014;2(11):901-10.
ND, Thompson CH, et 2 diabetes: systematic
30. Hollingsworth KG,
al. Comparison of low- review and meta-analysis
Abubacker MZ, Joubert
and high-carbohydrate of observational studies.
I, Allison MED, Lomas
diets for type 2 diabetes BMJ. 2015;351:h3978. DJ. Low-carbohydrate
management: a random- 25. Ericson U, Hellstrand diet induced reduction
ized trial. Am J Clin Nutr. S, Brunkwall L, et al. of hepatic lipid content
2015;102(4):780-90. Food sources of fat may observed with a rapid
21. Tay J, Brinkworth GD, clarify the inconsistent non-invasive MRI
Noakes M, et al. Metabolic role of dietary fat intake technique. Br J Radiol.
effects of weight loss on for incidence of type 2 2006;79:712–715.
a very-low-carbohydrate diabetes. Am J Clin Nutr. 31. Kaur B, Henry J. Micro-
diet compared with an 2015 May;101(5):1065-80. nutrient status in type
isocaloric high-carbohy- 26. Mayer SB, Jeffreys AS, 2 diabetes: a review. In:
drate diet in abdominally Olsen MK, et al. Two diets Henry J, editor. Advances
obese subjects. J Am Coll with different hemoglobin in food and nutrition
Cardiol. 2008;51(1):59-67. A1c and antiglycemic research. 71. Cambridge,
22. Siri-Tarino PW, Chiu S, medication effects despite Massachusetts: Academic
Bergeron N, Krauss RM. similar weight loss in type Press; 2014. p. 55-100.
73
NZMJ 1 April 2016, Vol 129 No 1432
ISSN 1175-8716 © NZMA
www.nzma.org.nz/journalYou can also read
