You are on page 1of 9

R E V I E W A R T I C L E

Type 2 Diabetes
Etiology and reversibility
ROY TAYLOR, MD, FRCP surgery and that 10 years later, almost
90% remained free of diabetes. The phe-
nomenon was more recently tested in a
randomized prospective study comparing

R
eversal of type 2 diabetes to normal Type 2 diabetes has long been known gastric banding with intensive medical
metabolic control by either bariat- to progress despite glucose-lowering therapy for type 2 diabetes (13). This least
ric surgery or hypocaloric diet allows treatment, with 50% of individuals re- invasive type of surgery was most suitable
for the time sequence of underlying path- quiring insulin therapy within 10 years for the randomized study, although it was
ophysiologic mechanisms to be observed. (1). This seemingly inexorable deteriora- associated with lower rates of diabetes re-
In reverse order, the same mechanisms are tion in control has been interpreted to versal than other procedures. Mean fast-
likely to determine the events leading to the mean that the condition is treatable but ing plasma glucose fell to normal levels in
onset of hyperglycemia and permit insight not curable. Clinical guidelines recognize the surgically treated group but declined
into the etiology of type 2 diabetes. Within this deterioration with algorithms of se- only modestly in the intensive medical
7 days of instituting a substantial negative quential addition of therapies. Insulin re- treatment group despite oral agents and
calorie balance by either dietary interven- sistance and b-cell dysfunction are insulin (Fig. 1) (13). Remission of diabe-
tion or bariatric surgery, fasting plasma known to be the major pathophysiologic tes was related to the degree of weight loss
glucose levels can normalize. This rapid factors driving type 2 diabetes; however, rather than to group allocation and was
change relates to a substantial fall in liver fat these factors come into play with very dif- achieved in 73% of the surgical group and
content and return of normal hepatic in- ferent time courses. Insulin resistance in 13% of the intensive medical treatment
sulin sensitivity. Over 8 weeks, rst phase muscle is the earliest detectable abnor- group because surgery was more effective
and maximal rates of insulin secretion mality of type 2 diabetes (2). In contrast, in achieving weight loss as previously
steadily return to normal, and this change changes in insulin secretion determine
is in step with steadily decreasing pancre- both the onset of hyperglycemia and the
atic fat content. The difference in time progression toward insulin therapy (3,4).
course of these two processes is striking. The etiology of each of these two major
Recent information on the intracellular factors appears to be distinct. Insulin re-
effects of excess lipid intermediaries ex- sistance may be caused by an insulin sig-
plains the likely biochemical basis, naling defect (5), glucose transporter
which simplies both the basic under- defect (6), or lipotoxicity (7), and b-cell
standing of the condition and the con- dysfunction is postulated to be caused by
cepts used to determine appropriate amyloid deposition in the islets (8), oxi-
management. Recent large, long-duration dative stress (9), excess fatty acid (10), or
population studies on time course lack of incretin effect (11). The demon-
of plasma glucose and insulin secretion stration of reversibility of type 2 diabetes
before the diagnosis of diabetes are con- offers the opportunity to evaluate the time
sistent with this new understanding. sequence of pathophysiologic events dur-
Type 2 diabetes has long been regarded ing return to normal glucose metabolism
as inevitably progressive, requiring in- and, hence, to unraveling the etiology.
creasing numbers of oral hypoglycemic
agents and eventually insulin, but it is Reversal of type 2 diabetes
now certain that the disease process can by bariatric surgerydThe rst
be halted with restoration of normal hint that type 2 diabetes is a fully re-
carbohydrate and fat metabolism. Type versible syndrome came from bariatric
2 diabetes can be understood as a poten- surgery. Almost a quarter century ago,
tially reversible metabolic state precipi- Pories et al. (12) demonstrated that blood
tated by the single cause of chronic glucose levels normalized in obese people
excess intraorgan fat. with type 2 diabetes undergoing bariatric Figure 1dA: Fasting plasma glucose and
c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c weight change 2 years after randomization
either to gastric banding or to intensive medi-
From the Magnetic Resonance Centre, Institute of Cellular Medicine, Newcastle University, Newcastle upon cal therapy for weight loss and glucose control.
Tyne, U.K.
Data plotted with permission from Dixon et al.
Corresponding author: Roy Taylor, roy.taylor@ncl.ac.uk.
Received 6 September 2012 and accepted 1 November 2012. (13). B: Early changes in fasting plasma glu-
DOI: 10.2337/dc12-1805 cose level following pancreatoduodenal bypass
2013 by the American Diabetes Association. Readers may use this article as long as the work is properly surgery. A decrease into the normal range was
cited, the use is educational and not for prot, and the work is not altered. See http://creativecommons.org/ seen within 7 days. Reproduced with permis-
licenses/by-nc-nd/3.0/ for details. sion from Taylor (98).

care.diabetesjournals.org DIABETES CARE, VOLUME 36, APRIL 2013 1047


Etiology of type 2 diabetes

described (14). Type 2 diabetes can be re- data following the dietary intervention
versed by applying a surgical procedure study are highly suggestive of a causal
that diminishes fat mass. link (21).
However, the observation that nor-
malization of glucose in type 2 diabetes New perspectives on insulin
occurred within days after bariatric sur- resistance
gery, before substantial weight loss (15),
led to the widespread belief that surgery Muscle
itself brought about specic changes me- Whole-body insulin resistance is the ear-
diated through incretin hormone secre- liest predictor of type 2 diabetes onset,
tion (16,17). This reasoning overlooked and this mainly reects muscle insulin
the major change that follows bariatric resistance (26). However, careful separa-
surgery: an acute, profound decrease in tion of the contributions of muscle and
calorie intake. Typically, those undergo- liver have shown that early improvement
ing bariatric surgery have a mean body in control of fasting plasma glucose level
weight of ;150 kg (15) and would there- is associated only with improvement in
fore require a daily calorie intake of liver insulin sensitivity (20,21). It is clear
;13.4 MJ/day (3,200 kcal/day) for that the resumption of normal or near-
weight maintenance (18). This intake de- normal diurnal blood glucose control
creases precipitously at the time of sur- does not require improvement in muscle
gery. The sudden reversal of trafc into insulin sensitivity. Although this nding
fat stores brings about a profound change may at rst appear surprising, it is sup-
in intracellular concentration of fat me- ported by a wide range of earlier observa-
tabolites. It is known that under hypo- tions. Mice totally lacking in skeletal
caloric conditions, fat is mobilized rst muscle insulin receptors do not develop
from the liver and other ectopic sites diabetes (27). Humans who have the
rather than from visceral or subcutaneous PPP1R3A genetic variant of muscle glyco-
fat stores (19). This process has been gen synthase cannot store glycogen in
studied in detail during more moderate muscle after meals but are not necessarily
calorie restriction in type 2 diabetes over hyperglycemic (28). Many normoglyce-
8 weeks (20). Fasting plasma glucose was Figure 2dEffect of a very-low-calorie diet in mic individuals maintain normal blood
shown to be improved because of an 81% type 2 diabetes on fasting plasma glucose level glucose levels with a degree of muscle in-
(A), basal hepatic glucose production (HGP)
decrease in liver fat content and normali- (B), and hepatic triacylglycerol content (C).
sulin resistance identical to those with
zation of hepatic insulin sensitivity with For comparison, data for a matched non- type 2 diabetes (29).
no change in the insulin resistance of diabetic control group are shown as . Re- Although a defect in mitochondrial
muscle. produced with permission from Lim et al. (21). function is associated with extremes of
FFM, fat-free mass. insulin resistance in skeletal muscle (30),
Reversal of type 2 diabetes this does not appear to be relevant to the
by diet alonedIf the rapid changes etiology of type 2 diabetes. No defect is
in metabolism following bariatric surgery (the rst phase insulin response) typically present in early type 2 diabetes but rather
are a consequence of the sudden change is absent. This was conrmed at baseline is directly related to ambient plasma glu-
in calorie balance, the defects in both in the study, but the rst phase response cose concentration (31). Observed rates
insulin secretion and hepatic insulin sen- increased gradually over 8 weeks of a of mitochondrial ATP production can be
sitivity of type 2 diabetes should be very-low-calorie diet to become indistin- modied by increasing or decreasing
correctable by change in diet alone. To guishable from that of age- and weight- plasma fatty acid concentration (32,33).
test this hypothesis, a group of people matched nondiabetic control subjects. Additionally, the onset of insulin stimula-
with type 2 diabetes were studied before The maximum insulin response, as elicited tion of mitochondrial ATP synthesis is
and during a 600 kcal/day diet (21). by arginine bolus during hyperglycemia, slow, gradually increasing over 2 h, and
Within 7 days, liver fat decreased by also normalized. Pancreas fat content quite distinct from the acute onset of in-
30%, becoming similar to that of the con- decreased gradually during the study pe- sulins metabolic effects (34). Although it
trol group, and hepatic insulin sensitivity riod to become the same as that in the remains possible that secondary mito-
normalized (Fig. 2). The close association control group, a time course matching chondrial effects of hyperglycemia and
between liver fat content and hepatic glu- that of the increase in both rst phase excess fatty acids exist, there is no evi-
cose production had previously been es- and total insulin secretion (Fig. 3). Fat dence for a primary mitochondrial defect
tablished (20,22,23). Plasma glucose content in the islets was not directly mea- underlying type 2 diabetes.
normalized by day 7 of the diet. sured, although it is known that islets take The physiologic importance of mus-
During this 8-week study, b-cell up fat avidly (24) and that islet fat content cle insulin resistance is likely to operate
function was tested by a gold standard closely reects total pancreatic fat content over a period of many years. The pres-
method that used a stepped glucose infu- in animal models (25). Although a cause- ence of long-standing muscle insulin
sion with subsequent arginine bolus (21). and-effect relationship between raised in- resistance will not of itself cause blood
In type 2 diabetes, the glucose-induced traorgan fat levels and metabolic effect glucose levels to rise, but raised plasma
initial rapid peak of insulin secretion has not yet been proven, the time course insulin levels will expedite accumulation

1048 DIABETES CARE, VOLUME 36, APRIL 2013 care.diabetesjournals.org


Taylor

levels, and both liver fat and serum ALT


returned to normal levels during a subse-
quent hypocaloric diet. Superimposed
upon a positive calorie balance, the extent
of portal vein hyperinsulinemia deter-
mines how rapidly conversion of excess
sugars to fatty acid occurs in the liver. In
groups of both obese and nonobese sub-
jects, it was found that those with higher
plasma insulin levels have markedly in-
creased rates of hepatic de novo lipogen-
esis (2,38,39). Conversely, in type 1
diabetes the relatively low insulin concen-
tration in the portal vein (as a con- Figure 4dMechanism of interaction between
sequence of insulin injection into excess amounts of fatty acids, diacylglycerol,
subcutaneous tissue) is associated with and ceramide and insulin action within the
subnormal liver fat content (40). Initiation hepatocyte. Diacylglycerol activates PKC
of subcutaneous insulin therapy in and inhibits activation of IRS-1 by the insulin
type 2 diabetes brings about a decrease receptor. Ceramides cause sequestration of
Akt2 by PKCz and inhibit insulin control of
in portal insulin delivery by suppression
gluconeogenesis. These mechanisms have re-
of pancreatic insulin secretion and, cently been reviewed (99). FFA, free-fatty
hence, a decrease in liver fat (41). Hypo- acid; TG, triacylglycerol.
caloric diet (42), physical activity (43), or
thiazolidinedione use (23,44) each re-
duces insulin secretion and decreases esteried to diacylglycerol, then PKC
liver fat content. Newly synthesized activation is prevented, and hepatic insulin
triacylglycerol in the liver will be either sensitivity is maintained. The molecular
oxidized, exported, or stored as hepatic specicity of this mechanism has been con-
triacylglycerol. Because transport of fatty rmed by use of antisense oligonucleotide
Figure 3dEffect of an 8-week very-low-
acid into mitochondria for oxidation is to PKC, which prevents hepatic insulin
calorie diet in type 2 diabetes on arginine-
induced maximal insulin secretion (A), rst inhibited by the malonyl-CoA produced resistance despite raised diacylglycerol
phase insulin response to a 2.8 mmol/L increase during de novo lipogenesis, newly synthe- levels during high-fat feeding (47). In obese
in plasma glucose (B), and pancreas triacyl- sized triacylglycerol is preferentially di- humans, intrahepatic diacylglycerol con-
glycerol (TG) content (C). For comparison, rected toward storage or export. Hence, centration has been shown to correlate
data for a matched nondiabetic control group are hepatic fat content and plasma VLDL tri- with hepatic insulin sensitivity (48,49).
shown as . Replotted with permission from Lim acylglycerol levels are increased. Additionally, the presence of excess fatty
et al. (21). Within the hepatocyte, fatty acids can acids promotes ceramide synthesis by es-
only be derived from de novo lipogenesis, terication with sphingosine. Ceramides
of liver fat by stimulation of de novo uptake of nonesteried fatty acid and LDL, cause sequestration of Akt2 and activation
lipogenesis (26). or lipolysis of intracellular triacylglycerol. of gluconeogenic enzymes (Fig. 4), al-
The fatty acid pool may be oxidized for though no relationship with in vivo
Liver energy or may be combined with glycerol to insulin resistance could be demonstrated
Evidence linking hepatic insulin sensitiv- form mono-, di-, and then triacylglycerols. in humans (49). However, the described in-
ity to intraorgan triglyceride content has It is possible that a lower ability to oxidize tracellular regulatory roles of diacylglycerol
been steadily accumulating. In insulin- fat within the hepatocyte could be one of and ceramide are consistent with the in
treated type 2 diabetes, insulin dose corre- several susceptibility factors for the accu- vivo observations of hepatic steatosis and
lates with the extent of fatty liver (35), and mulation of liver fat (45). Excess diacylgly- control of hepatic glucose production
in turn, this is associated with insulin sen- cerol has a profound effect on activating (20,21).
sitivity to suppression of hepatic glucose protein kinase C epsilon type (PKC), Fasting plasma glucose concentration
production (36). Decreasing the fat content which inhibits the signaling pathway depends entirely on the fasting rate of
of liver is associated with improvement in from the insulin receptor to insulin recep- hepatic glucose production and, hence,
insulin suppression of glucose production tor substrate 1 (IRS-1), the rst postrecep- on its sensitivity to suppression by in-
and, thereby, with improvement in fasting tor step in intracellular insulin action (46). sulin. Hepatic insulin sensitivity cannot
plasma glucose (20,23). Thus, under circumstances of chronic en- be inferred from observed postprandial
Storage of liver fat can only occur ergy excess, a raised level of intracellular change in hepatic glycogen concentration
when daily calorie intake exceeds expen- diacylglycerol specically prevents normal because glucose transport into the hepato-
diture. Sucrose overfeeding for 3 weeks insulin action, and hepatic glucose produc- cyte is not rate limiting, unlike in muscle,
has been shown to cause a 30% increase tion fails to be controlled (Fig. 4). High-fat and hyperglycemia itself drives the pro-
in liver fat content (37). The associated feeding of rodents brings about raised lev- cess of glycogen synthesis irrespective of
metabolic stress on hepatocytes was re- els of diacylglycerol, PKC activation, and insulin action. Indeed, postprandial gly-
ected by a simultaneous 30% rise in se- insulin resistance. However, if fatty acids cogen storage in liver has been shown to
rum alanine aminotransferase (ALT) are preferentially oxidized rather than be moderately impaired in type 2 diabetes

care.diabetesjournals.org DIABETES CARE, VOLUME 36, APRIL 2013 1049


Etiology of type 2 diabetes

(50) compared with the marked impair- and p18, and this effect is magnied by level of 5%? There must be varying de-
ment in skeletal muscle (51). increased glucose concentration (64). grees of liposusceptibility of the metabolic
Although a close relationship exists This antiproliferative effect is specically organs, and this has been demonstrated in
among raised liver fat levels, insulin re- prevented by small interfering RNA relation to ethnic differences (72). If the fat
sistance, and raised liver enzyme levels knockdown of the inhibitors. In the is simply not available to the body, then
(52), high levels of liver fat are not inevi- Zucker diabetic fatty rat, a genetic model the susceptibility of the pancreas will not
tably associated with hepatic insulin of spontaneous type 2 diabetes, the onset be tested, whereas if the individual ac-
resistance. This is analogous to the dis- of hyperglycemia is preceded by a rapid quires excess fat stores, then b-cell failure
cordance observed in the muscle of increase in pancreatic fat (58). It is partic- may or may not develop depending on de-
trained athletes in whom raised intramyo- ularly noteworthy that the onset of diabe- gree of liposusceptibility. In any group of
cellular triacylglycerol is associated with tes in this genetic model is completely people with type 2 diabetes, simple in-
high insulin sensitivity (53). This rela- preventable by restriction of food intake spection reveals that diabetes develops in
tionship is also seen in muscle of mice (65), illustrating the interaction between some with a body mass index (BMI) in the
overexpressing the enzyme DGAT-1, genetic susceptibility and environmental normal or overweight range, whereas oth-
which rapidly esteries diacylglycerol to factors. ers have a very high BMI. The pathophys-
metabolically inert triacylglycerol (54). In Clearly separate from the characteris- iologic changes in insulin secretion and
both circumstances, raised intracellular tic lack of acute insulin secretion in re- insulin sensitivity are not different in
triacylglycerol stores coexist with normal sponse to increase in glucose supply is the obese and normal weight people (73),
insulin sensitivity. When a variant of matter of total mass of b-cells. The former and the upswing in population rates of
PNPLA3 was described as determining in- determines the immediate metabolic re- type 2 diabetes relates to a right shift in
creased hepatic fat levels, it appeared sponse to eating, whereas the latter the whole BMI distribution. Hence, the
that a major factor underlying nonalco- places a long-term limitation on total pos- person with a BMI of 24 and type 2 diabe-
holic fatty liver disease and insulin resis- sible insulin response. Histological stud- tes would in a previous era have had a BMI
tance was identied (55). However, this ies of the pancreas in type 2 diabetes of 21 and no diabetes. It is clear that in-
relatively rare genetic variant is not asso- consistently show an ;50% reduction dividual susceptibility factors determine
ciated with hepatic insulin resistance in number of b-cells compared with nor- the onset of the condition, and both ge-
(56). Because the responsible G allele of mal subjects (66). b-Cell loss appears to netic and epigenetic factors may contrib-
PNPLA3 is believed to code for a lipase increase as duration of diabetes increases ute. Given that diabetes cannot occur
that is ineffective in triacylglycerol hydro- (67). The process is likely to be regulated without loss of acute insulin response to
lysis, it appears that diacylglycerol and by apoptosis, a mechanism known to be food, it can be postulated that this failure
fatty acids are sequestered as inert triacyl- increased by chronic exposure to in- of acute insulin secretion could relate to
glycerol, preventing any inhibitory effect creased fatty acid metabolites (68). both accumulation of fat and susceptibil-
on insulin signaling. Ceramides, which are synthesized directly ity to the adverse effect of excess fat in the
from fatty acids, are likely mediators of pancreas.
New perspectives on the the lipid effects on apoptosis (10,69). In
b-cell defectdChronic exposure of light of new knowledge about b-cell apo- The time course to
b-cells to triacylglycerol or fatty acids ei- ptosis and rates of turnover during adult development of type 2
ther in vitro or in vivo decreases b-cell life, it is conceivable that removal of ad- diabetesdThe earliest predictor of
capacity to respond to an acute increase verse factors could result in restoration of the development of type 2 diabetes is
in glucose levels (57,58). This concept is normal b-cell number, even late in the dis- low insulin sensitivity in skeletal muscle,
far from new (59,60), but the observa- ease (66,70). Plasticity of lineage and but it is important to recognize that this is
tions of what happens during reversal of transdifferentiation of human adult not a distinct abnormality but rather part
diabetes provide a new perspective. b-cells could also be relevant, and the ev- of the wide range expressed in the pop-
b-Cells avidly import fatty acids through idence for this has recently been reviewed ulation. Those people in whom diabetes
the CD36 transporter (24,61) and re- (71). b-Cell number following reversal of will develop simply have insulin sensitiv-
spond to increased fatty acid supply by type 2 diabetes remains to be examined, ity, mainly in the lowest population quar-
storing the excess as triacylglycerol (62). but overall, it is clear that at least a critical tile (29). In prediabetic individuals, raised
The cellular process of insulin secretion in mass of b-cells is not permanently dam- plasma insulin levels compensate and al-
response to an increase in glucose supply aged but merely metabolically inhibited. low normal plasma glucose control. How-
depends on ATP generation by glucose A wide scatter of absolute levels of ever, because the process of de novo
oxidation. However, in the context of an pancreas triacylglycerol has been reported, lipogenesis is stimulated by higher insulin
oversupply of fatty acids, such chronic with a tendency for higher levels in levels (38), the scene is set for hepatic fat
nutrient surfeit prevents further increases people with diabetes (57). This large pop- accumulation. Excess fat deposition in the
in ATP production. Increased fatty acid ulation study showed overlap between di- liver is present before the onset of classical
availability inhibits both pyruvate cy- abetic and weight-matched control type 2 diabetes (43,7476), and in estab-
cling, which is normally increased during groups. These ndings were also ob- lished type 2 diabetes, liver fat is supra-
an acute increase in glucose availability, served in a more recent smaller study normal (20). When ultrasound rather
and pyruvate dehydrogenase activity, the that used a more precise method (21). than magnetic resonance imaging is
major rate-limiting enzyme of glucose ox- Why would one person have normal used, only more-severe degrees of steato-
idation (63). Fatty acids have been shown b-cell function with a pancreas fat level sis are detected, and the prevalence of
to inhibit b-cell proliferation in vitro by of, for example, 8%, whereas another fatty liver is underestimated, with esti-
induction of the cell cycle inhibitors p16 has type 2 diabetes with a pancreas fat mates of 70% of people with type 2

1050 DIABETES CARE, VOLUME 36, APRIL 2013 care.diabetesjournals.org


Taylor

diabetes as having a fatty liver (76). None- diabetes (Fig. 6): The accumulation of
theless, the prognostic power of merely fat in liver and secondarily in the pancreas
the presence of a fatty liver is impressive will lead to self-reinforcing cycles that in-
of predicting the onset of type 2 diabetes. teract to bring about type 2 diabetes. Fatty
A large study of individuals with normal liver leads to impaired fasting glucose me-
glucose tolerance at baseline showed a tabolism and increases export of VLDL
very low 8-year incidence of type 2 diabe- triacylglycerol (85), which increases fat
tes if fatty liver had been excluded at base- delivery to all tissues, including the islets.
line, whereas if present, the hazard ratio The liver and pancreas cycles drive on-
for diabetes was 5.5 (range 3.68.5) (74). ward after diagnosis with steadily de-
In support of this nding, a temporal pro- creasing b-cell function. However, of
gression from weight gain to raised liver note, observations of the reversal of type 2
enzyme levels and onward to hypertrigly- diabetes conrm that if the primary inu-
ceridemia and then glucose intolerance ence of positive calorie balance is removed,
has been demonstrated (77). then the processes are reversible (21).
In obese young people, decreased How long will diabetes stay away after
b-cell function has recently been shown weight loss? Long-term normal blood
to predict deterioration of glucose toler- glucose control in previously diabetic
ance (4,78). Additionally, the rate of de- individuals after bariatric surgery demon-
cline in glucose tolerance in rst-degree strates that diabetes does not recur for up
relatives of type 2 diabetic individuals is to 10 years, unless substantial weight gain
strongly related to the loss of b-cell func- occurs (86). These observations are con-
tion, whereas insulin sensitivity changes sistent with the twin cycle hypothesis and
little (79). This observation mirrors those the existence of a trigger level for adverse
in populations with a high incidence of metabolic effects of fat in the pancreas.
type 2 diabetes in which transition from Hence, for a given individual with type
hyperinsulinemic normal glucose toler- 2 diabetes, reducing the liver and pan-
ance to overt diabetes involves a large, creas fat content below his or her personal
rapid rise in glucose levels as a result trigger levels would be expected to result
of a relatively small further loss of acute in a release from the fatty acidmediated
b-cell competence (3). The Whitehall II Figure 5dChange in fasting plasma glucose dysfunction. Individual tolerance of dif-
study showed in a large population fol- (A), 2 h post-oral glucose tolerance test (B), ferent degrees of fat exposure vary, and
lowed prospectively that people with di- and homeostasis model assessment (HOMA-B) understanding this liposusceptibility will
abetes exhibit a sudden rise in fasting insulin secretion (C) during the 16-year follow- underpin the future understanding of ge-
up in the Whitehall II study. Of the 6,538 people
glucose as b-cell function deteriorates netically determined risk in any given en-
studied, diabetes developed in 505. Time 0 was
(Fig. 5) (80). Hence, the ability of the pan- taken as the diagnosis of diabetes or as the end of vironment. However, this should not
creas to mount a normal, brisk insulin re- follow-up for those remaining normoglycemic. obscure the central point: If a person has
sponse to an increasing plasma glucose Redrawn with permission from Tabak et al. type 2 diabetes, there is more fat in the
level is lost in the 2 years before the de- (80). liver and pancreas than he or she can cope
tection of diabetes, although fasting with.
plasma glucose levels may have been at
the upper limit of normal for several clear that the b-cell defect, not solely he- Implication for management
years. This was very different from the patic insulin resistance, may be reversible of type 2 diabetesdThe extent of
widely assumed linear rise in fasting by weight loss at least early in the course of weight loss required to reverse type 2
plasma glucose level and gradual b-cell type 2 diabetes (21,84). The low insulin diabetes is much greater than conven-
decompensation but is consistent with sensitivity of muscle tissue does not tionally advised. A clear distinction must
the time course of markers of increased change materially either during the onset be made between weight loss that im-
liver fat before the onset of type 2 diabetes of diabetes or during subsequent reversal. proves glucose control but leaves blood
observed in other studies (81). Data from Overall, the information on the inhibitory glucose levels abnormal and weight loss
the West of Scotland Coronary Preven- effects of excess fat on b-cell function and of sufcient degree to normalize pancre-
tion Study demonstrated that plasma apoptosis permits a new understanding atic function. The Belfast diet study pro-
triacylglycerol and ALT levels were mod- of the etiology and time course of type 2 vides an example of moderate weight loss
estly elevated 2 years before the diagnosis diabetes. leading to reasonably controlled, yet per-
of type 2 diabetes and that there was a sistent diabetes. This study showed that a
steady rise in the level of this liver The twin cycle hypothesis mean weight loss of 11 kg decreased
enzyme in the run-up to the time of di- of etiology of type 2 fasting blood glucose levels from 10.4 to
agnosis (75). diabetesdTogether with evidence of 7.0 mmol/L but that this abnormal level
The accepted view has been that the normalization of insulin secretion after presaged the all-too-familiar deterioration
b-cell dysfunction of established dia- bariatric surgery (84), insights into the of control (87).
betes progresses inexorably (79,82,83), behavior of the liver and pancreas during Data from the Swedish randomized
whereas insulin resistance can be modied hypocaloric dieting lead to a hypothesis of study of gastric banding showed that a
at least to some extent. However, it is now the etiology and pathogenesis of type 2 loss of 20% body weight was associated

care.diabetesjournals.org DIABETES CARE, VOLUME 36, APRIL 2013 1051


Etiology of type 2 diabetes

alone (93). However, exercise programs


alone produce no weight loss for over-
weight middle-aged people (94). The nec-
essary initial major loss of body weight
demands a substantial reduction in en-
ergy intake. After weight loss, steady
weight is most effectively achieved by a
combination of dietary restriction and
physical activity. Both aerobic and resis-
tance exercise are effective (95). The crit-
ical factor is sustainability.
Formal recommendations on how to
reverse type 2 diabetes in clinical practice
must await further studies. In the mean-
time, it will be helpful for all individuals
with newly diagnosed type 2 diabetes to
know that they have a metabolic syndrome
that is reversible. They should know that if
it is not reversed, the consequences for
future health and cost of life insurance are
dire, although these serious adverse effects
must be balanced against the difculties
and privations associated with a substantial
and sustained change in eating patterns.
For many people, this may prove to be too
Figure 6dThe twin cycle hypothesis of the etiology of type 2 diabetes. During long-term intake of high a price to pay, but for those who are
more calories than are expended each day, any excess carbohydrate must undergo de novo li- strongly motivated to escape from type 2
pogenesis, which particularly promotes fat accumulation in the liver. Because insulin stimulates diabetes, the new understanding gives clear
de novo lipogenesis, individuals with a degree of insulin resistance (determined by family or direction. Physicians need to accept that
lifestyle factors) will accumulate liver fat more readily than others because of higher plasma long-term weight loss is achievable for a
insulin levels. In turn, the increased liver fat will cause relative resistance to insulin suppression of
hepatic glucose production. Over many years, a modest increase in fasting plasma glucose level
worthwhile proportion of patients (96). In
will stimulate increased basal insulin secretion rates to maintain euglycemia. The consequent the United States, diabetes costs $174 bil-
hyperinsulinemia will further increase the conversion of excess calories to liver fat. A cycle of lion annually (97), and in the United King-
hyperinsulinemia and blunted suppression of hepatic glucose production becomes established. dom, it accounts for 10% of National
Fatty liver leads to increased export of VLDL triacylglycerol (85), which will increase fat delivery Health Service expenditure. Even if only a
to all tissues, including the islets. This process is further stimulated by elevated plasma glucose small proportion of patients with type 2
levels (85). Excess fatty acid availability in the pancreatic islet would be expected to impair the diabetes return to normal glucose control,
acute insulin secretion in response to ingested food, and at a certain level of fatty acid exposure, the savings in disease burden and eco-
postprandial hyperglycemia will supervene. The hyperglycemia will further increase insulin nomic cost will be enormous.
secretion rates, with consequent enhancement of hepatic lipogenesis, spinning the liver cycle
faster and driving the pancreas cycle. Eventually, the fatty acid and glucose inhibitory effects on
the islets reach a trigger level that leads to a relatively sudden onset of clinical diabetes. Figure AcknowledgmentsdThe research was sup-
adapted with permission from Taylor (98). ported by the National Institute for Health Re-
search Newcastle Biomedical Research Centre.
The funder played no role in the conduct
of the study, collection of data, management of
with long-term remission in 73% of a especially the Ala12 allele at PPARG, of the study, analysis of data, interpretation of
bariatric surgery group, with weight successful long-term weight loss have data, or preparation of the manuscript.
change itself being the principal determi- been identied (89), and use of such No potential conicts of interest relevant to
nant of glucose control (13). Dietary markers could guide future therapy. It this article were reported.
weight loss of 15 kg allowed for reversal must be noted that involuntary food Parts of this study were presented by R.T. at
of diabetes in a small group of individuals shortage, such as a result of war, results 2012 Banting Memorial Lecture at the Annual
recently receiving a diagnosis (21). In in- in a sharp fall in type 2 diabetes prevalence Professional Conference of Diabetes UK,
dividuals strongly motivated to regain (90,91). Glasgow, U.K., 79 March 2012.
The author gratefully acknowledges the
normal health, substantial weight loss is The role of physical activity must be valuable comments of Prof. Sally Marshall
entirely possible by decreasing food con- considered. Increased levels of daily ac- (Newcastle University) and Prof. John Simpson
sumption (88). This information should tivity bring about decreases in liver fat (Newcastle University) on the manuscript.
be made available to all people with type 2 stores (43), and a single bout of exercise
diabetes, even though with present meth- substantially decreases both de novo lipo-
ods of changing eating habits, it is un- genesis (39) and plasma VLDL (92). Sev- References
likely that weight loss can be achieved in eral studies demonstrated that calorie 1. Turner RC, Cull CA, Frighi V, Holman
those not strongly motivated to escape control combined with exercise is much RR; UK Prospective Diabetes Study
from diabetes. Some genetic predictors, more successful than calorie restriction (UKPDS) Group. Glycemic control with

1052 DIABETES CARE, VOLUME 36, APRIL 2013 care.diabetesjournals.org


Taylor

diet, sulfonylurea, metformin, or insulin 15. Guidone C, Manco M, Valera-Mora E, knockout exhibits features of the meta-
in patients with type 2 diabetes mellitus: et al. Mechanisms of recovery from type 2 bolic syndrome of NIDDM without alter-
progressive requirement for multiple diabetes after malabsorptive bariatric ing glucose tolerance. Mol Cell 1998;2:
therapies (UKPDS 49). JAMA 1999;281: surgery. Diabetes 2006;55:20252031 559569
20052012 16. Rubino F, Forgione A, Cummings DE, 28. Savage DB, Zhai L, Ravikumar B, et al.
2. Petersen KF, Dufour S, Morino K, Yoo PS, et al. The mechanism of diabetes control A prevalent variant in PPP1R3A impairs
Cline GW, Shulman GI. Reversal of mus- after gastrointestinal bypass surgery reveals a glycogen synthesis and reduces muscle
cle insulin resistance by weight reduction role of the proximal small intestine in the glycogen content in humans and mice
in young, lean, insulin-resistant offspring pathophysiology of type 2 diabetes. Ann [published correction in: Plos Med 2008;
of parents with type 2 diabetes. Proc Natl Surg 2006;244:741749 5:e246]. PLoS Med 2008;5:e27
Acad Sci U S A 2012;109:82368240 17. Holst JJ. Postprandial insulin secretion 29. Taylor R. Insulin resistance and type 2
3. Ferrannini E, Nannipieri M, Williams K, after gastric bypass surgery: the role of diabetes. Diabetes 2012;61:778779
Gonzales C, Haffner SM, Stern MP. Mode glucagon-like peptide 1. Diabetes 2011; 30. Petersen KF, Dufour S, Befroy D, Garcia R,
of onset of type 2 diabetes from normal or 60:22032205 Shulman GI. Impaired mitochondrial ac-
impaired glucose tolerance. Diabetes 18. Gibney ER, Murgatroyd P, Wright A, Jebb S, tivity in the insulin-resistant offspring of
2004;53:160165 Elia M. Measurement of total energy ex- patients with type 2 diabetes. N Engl
4. Cali AM, Man CD, Cobelli C, et al. Primary penditure in grossly obese women: com- J Med 2004;350:664671
defects in beta-cell function further exac- parison of the bicarbonate-urea method 31. Schrauwen-Hinderling VB, Kooi ME,
erbated by worsening of insulin resistance with whole-body calorimetry and free- Hesselink MK, et al. Impaired in vivo
mark the development of impaired glu- living doubly labelled water. Int J Obes Relat mitochondrial function but similar intra-
cose tolerance in obese adolescents. Di- Metab Disord 2003;27:641647 myocellular lipid content in patients with
abetes Care 2009;32:456461 19. Colles SL, Dixon JB, Marks P, Strauss BJ, type 2 diabetes mellitus and BMI-matched
5. Cusi K, Maezono K, Osman A, et al. In- OBrien PE. Preoperative weight loss control subjects. Diabetologia 2007;50:
sulin resistance differentially affects the PI with a very-low-energy diet: quantitation 113120
3-kinase- and MAP kinase-mediated sig- of changes in liver and abdominal fat by 32. Brehm A, Krssak M, Schmid AI, Nowotny
naling in human muscle. J Clin Invest serial imaging. Am J Clin Nutr 2006;84: P, Waldhusl W, Roden M. Increased
2000;105:311320 304311 lipid availability impairs insulin-stimulated
6. Herman MA, Kahn BB. Glucose transport 20. Petersen KF, Dufour S, Befroy D, Lehrke M, ATP synthesis in human skeletal muscle.
and sensing in the maintenance of glucose Hendler RE, Shulman GI. Reversal of non- Diabetes 2006;55:136140
homeostasis and metabolic harmony. alcoholic hepatic steatosis, hepatic insulin 33. Lim EL, Hollingsworth KG, Smith FE,
J Clin Invest 2006;116:17671775 resistance, and hyperglycemia by moderate Thelwall PE, Taylor R. Inhibition of li-
7. Roden M, Price TB, Perseghin G, et al. weight reduction in patients with type 2 polysis in Type 2 diabetes normalizes
Mechanism of free fatty acid-induced in- diabetes. Diabetes 2005;54:603608 glucose disposal without change in muscle
sulin resistance in humans. J Clin Invest 21. Lim EL, Hollingsworth KG, Aribisala BS, glycogen synthesis rates. Clin Sci (Lond)
1996;97:28592865 Chen MJ, Mathers JC, Taylor R. Reversal 2011;121:169177
8. Hull RL, Westermark GT, Westermark P, of type 2 diabetes: normalisation of beta 34. Lim EL, Hollingsworth KG, Thelwall PE,
Kahn SE. Islet amyloid: a critical entity in cell function in association with decreased Taylor R. Measuring the acute effect of in-
the pathogenesis of type 2 diabetes. J Clin pancreas and liver triacylglycerol. Dia- sulin infusion on ATP turnover rate in hu-
Endocrinol Metab 2004;89:36293643 betologia 2011;54:25062514 man skeletal muscle using phosphorus-31
9. Roma LP, Pascal SM, Duprez J, Jonas JC. 22. Gastaldelli A, Cusi K, Pettiti M, et al. Re- magnetic resonance saturation transfer
Mitochondrial oxidative stress contrib- lationship between hepatic/visceral fat spectroscopy. NMR Biomed 2010;23:952
utes differently to rat pancreatic islet cell and hepatic insulin resistance in non- 957
apoptosis and insulin secretory defects diabetic and type 2 diabetic subjects. 35. Ryysy L, Hkkinen A-M, Goto T, et al.
after prolonged culture in a low non- Gastroenterology 2007;133:496506 Hepatic fat content and insulin action on
stimulating glucose concentration. Dia- 23. Ravikumar B, Gerrard J, Dalla Man C, et al. free fatty acids and glucose metabolism
betologia 2012;55:22262237 Pioglitazone decreases fasting and post- rather than insulin absorption are associ-
10. Cnop M. Fatty acids and glucolipotoxicity prandial endogenous glucose production ated with insulin requirements during
in the pathogenesis of Type 2 diabetes. in proportion to decrease in hepatic tri- insulin therapy in type 2 diabetic patients.
Biochem Soc Trans 2008;36:348352 glyceride content. Diabetes 2008;57: Diabetes 2000;49:749758
11. Ahrn B, Gomis R, Standl E, Mills D, 22882295 36. Seppl-Lindroos A, Vehkavaara S,
Schweizer A. Twelve- and 52-week ef- 24. Lalloyer F, Vandewalle B, Percevault F, Hkkinen AM, et al. Fat accumulation in
cacy of the dipeptidyl peptidase IV in- et al. Peroxisome proliferator-activated the liver is associated with defects in in-
hibitor LAF237 in metformin-treated receptor alpha improves pancreatic ad- sulin suppression of glucose production
patients with type 2 diabetes. Diabetes aptation to insulin resistance in obese and serum free fatty acids independent of
Care 2004;27:28742880 mice and reduces lipotoxicity in human obesity in normal men. J Clin Endocrinol
12. Pories WJ, MacDonald KG Jr, Morgan EJ, islets. Diabetes 2006;55:16051613 Metab 2002;87:30233028
et al. Surgical treatment of obesity and its 25. Lee Y, Lingvay I, Szczepaniak LS, Ravazzola 37. Sevastianova K, Santos A, Kotronen A,
effect on diabetes: 10-y follow-up. Am J M, Orci L, Unger RH. Pancreatic steatosis: et al. Effect of short-term carbohydrate
Clin Nutr 1992;55(Suppl.):582S585S harbinger of type 2 diabetes in obese ro- overfeeding and long-term weight loss on
13. Dixon JB, OBrien PE, Playfair J, et al. dents. Int J Obes (Lond) 2010;34:396 liver fat in overweight humans. Am J Clin
Adjustable gastric banding and conven- 400 Nutr 2012;96:727734
tional therapy for type 2 diabetes: a ran- 26. Petersen KF, Dufour S, Savage DB, et al. 38. Schwarz JM, Linfoot P, Dare D, Aghajanian
domized controlled trial. JAMA 2008; The role of skeletal muscle insulin re- K. Hepatic de novo lipogenesis in nor-
299:316323 sistance in the pathogenesis of the meta- moinsulinemic and hyperinsulinemic sub-
14. Ferrannini E, Mingrone G. Impact of dif- bolic syndrome. Proc Natl Acad Sci U S A jects consuming high-fat, low-carbohydrate
ferent bariatric surgical procedures on in- 2007;104:1258712594 and low-fat, high-carbohydrate iso-
sulin action and beta-cell function in type 2 27. Bruning JC, Michael MD, Winnay JN, energetic diets. Am J Clin Nutr 2003;77:
diabetes. Diabetes Care 2009;32:514520 et al. A muscle-specic insulin receptor 4350

care.diabetesjournals.org DIABETES CARE, VOLUME 36, APRIL 2013 1053


Etiology of type 2 diabetes

39. Rabl R, Petersen KF, Dufour S, Flannery C, pathological study and effect of lifestyle 66. Butler AE, Janson J, Bonner-Weir S, Ritzel
Shulman GI. Reversal of muscle insulin re- advice. Hepatology 2006;44:458465 R, Rizza RA, Butler PC. Beta-cell decit
sistance with exercise reduces postprandial 53. van Loon LJ, Goodpaster BH. Increased and increased beta-cell apoptosis in hu-
hepatic de novo lipogenesis in insulin re- intramuscular lipid storage in the insulin- mans with type 2 diabetes. Diabetes 2003;
sistant individuals. Proc Natl Acad Sci U S A resistant and endurance-trained state. 52:102110
2011;108:1370513709 Pugers Arch 2006;451:606616 67. Rahier J, Guiot Y, Goebbels RM, Sempoux
40. Perseghin G, Lattuada G, De Cobelli F, 54. Liu L, Shi X, Choi CS, et al. Paradoxical C, Henquin JC. Pancreatic beta-cell mass
et al. Reduced intrahepatic fat content is coupling of triglyceride synthesis and fatty in European subjects with type 2 diabetes.
associated with increased whole-body acid oxidation in skeletal muscle over- Diabetes Obes Metab 2008;10(Suppl. 4):
lipid oxidation in patients with type expressing DGAT1. Diabetes 2009;58: 3242
1 diabetes. Diabetologia 2005;48:2615 25162524 68. Shimabukuro M, Higa M, Zhou YT, Wang
2621 55. Romeo S, Kozlitina J, Xing C, et al. Genetic MY, Newgard CB, Unger RH. Lipo-
41. Juurinen L, Tiikkainen M, Hkkinen AM, variation in PNPLA3 confers susceptibility apoptosis in beta-cells of obese prediabetic
Hakkarainen A, Yki-Jrvinen H. Effects of to nonalcoholic fatty liver disease. Nat fa/fa rats. Role of serine palmitoyltransferase
insulin therapy on liver fat content and Genet 2008;40:14611465 overexpression. J Biol Chem 1998;273:
hepatic insulin sensitivity in patients with 56. Kantartzis K, Peter A, Machicao F, et al. 3248732490
type 2 diabetes. Am J Physiol Endocrinol Dissociation between fatty liver and in- 69. Shimabukuro M, Zhou YT, Levi M, Unger
Metab 2007;292:E829E835 sulin resistance in humans carrying a RH. Fatty acid-induced beta cell apopto-
42. Nobili V, Marcellini M, Marchesini G, variant of the patatin-like phospholipase 3 sis: a link between obesity and diabetes.
et al. Intrauterine growth retardation, in- gene. Diabetes 2009;58:26162623 Proc Natl Acad Sci U S A 1998;95:2498
sulin resistance, and nonalcoholic fatty 57. Tushuizen ME, Bunck MC, Pouwels PJ, 2502
liver disease in children. Diabetes Care et al. Pancreatic fat content and beta-cell 70. Butler AE, Cao-Minh L, Galasso R, et al.
2007;30:26382640 function in men with and without type 2 Adaptive changes in pancreatic beta cell
43. Perseghin G, Lattuada G, De Cobelli F, diabetes. Diabetes Care 2007;30:2916 fractional area and beta cell turnover in
et al. Habitual physical activity is associ- 2921 human pregnancy. Diabetologia 2010;53:
ated with intrahepatic fat content in hu- 58. Lee Y, Hirose H, Ohneda M, Johnson JH, 21672176
mans. Diabetes Care 2007;30:683688 McGarry JD, Unger RH. Beta-cell lipo- 71. Szabat M, Lynn FC, Hoffman BG, Kieffer TJ,
44. Belfort R, Harrison SA, Brown K, et al. toxicity in the pathogenesis of non-insulin- Allan DW, Johnson JD. Maintenance of
A placebo-controlled trial of pioglitazone dependent diabetes mellitus of obese rats: b-cell maturity and plasticity in the adult
in subjects with nonalcoholic steatohe- impairment in adipocyte-beta-cell relation- pancreas: developmental biology concepts
patitis. N Engl J Med 2006;355:2297 ships. Proc Natl Acad Sci U S A 1994;91: in adult physiology. Diabetes 2012;61:
2307 1087810882 13651371
45. Szendroedi J, Chmelik M, Schmid AI, 59. McGarry JD. Banting lecture 2001: dys- 72. Szczepaniak LS, Victor RG, Mathur R,
et al. Abnormal hepatic energy homeo- regulation of fatty acid metabolism in the et al. Pancreatic steatosis and its relation-
stasis in type 2 diabetes. Hepatology etiology of type 2 diabetes. Diabetes 2002; ship to b-cell dysfunction in humans: ra-
2009;50:10791086 51:718 cial and ethnic variations. Diabetes Care
46. Samuel VT, Petersen KF, Shulman GI. 60. Unger RH. Lipotoxicity in the pathogen- 2012;35:23772383
Lipid-induced insulin resistance: un- esis of obesity-dependent NIDDM. Ge- 73. Reaven GM, Doberne L, Greeneld MS.
ravelling the mechanism. Lancet 2010; netic and clinical implications. Diabetes Comparison of insulin secretion and in
375:22672277 1995;44:863870 vivo insulin action in nonobese and
47. Savage DB, Choi CS, Samuel VT, et al. 61. Noushmehr H, DAmico E, Farilla L, et al. moderately obese individuals with non-
Reversal of diet-induced hepatic steatosis Fatty acid translocase (FAT/CD36) is lo- insulin-dependent diabetes mellitus. Di-
and hepatic insulin resistance by antisense calized on insulin-containing granules in abetes 1982;31:382384
oligonucleotide inhibitors of acetyl-CoA human pancreatic beta-cells and mediates 74. Shibata M, Kihara Y, Taguchi M, Tashiro M,
carboxylases 1 and 2. J Clin Invest 2006; fatty acid effects on insulin secretion. Di- Otsuki M. Nonalcoholic fatty liver disease
116:817824 abetes 2005;54:472481 is a risk factor for type 2 diabetes in middle-
48. Magkos F, Su X, Bradley D, et al. Intra- 62. Diakogiannaki E, Dhayal S, Childs CE, aged Japanese men. Diabetes Care 2007;30:
hepatic diacylglycerol content is associ- Calder PC, Welters HJ, Morgan NG. 29402944
ated with hepatic insulin resistance in Mechanisms involved in the cytotoxic and 75. Sattar N, McConnachie A, Ford I, et al.
obese subjects. Gastroenterology 2012; cytoprotective actions of saturated versus Serial metabolic measurements and con-
142:14441446 monounsaturated long-chain fatty acids version to type 2 diabetes in the west of
49. Kumashiro N, Erion DM, Zhang D, et al. in pancreatic beta-cells. J Endocrinol Scotland coronary prevention study:
Cellular mechanism of insulin resistance 2007;194:283291 specic elevations in alanine aminotrans-
in nonalcoholic fatty liver disease. Proc 63. Zhou YP, Priestman DA, Randle PJ, Grill ferase and triglycerides suggest hepatic fat
Natl Acad Sci U S A 2011;108:16381 VE. Fasting and decreased B cell sensitiv- accumulation as a potential contributing
16385 ity: important role for fatty acid-induced factor. Diabetes 2007;56:984991
50. Krssak M, Brehm A, Bernroider E, et al. inhibition of PDH activity. Am J Physiol 76. Targher G, Bertolini L, Padovani R, et al.
Alterations in postprandial hepatic gly- 1996;270:E988E994 Prevalence of nonalcoholic fatty liver dis-
cogen metabolism in type 2 diabetes. Di- 64. Pascoe J, Hollern D, Stamateris R, et al. ease and its association with cardiovas-
abetes 2004;53:30483056 Free fatty acids block glucose-induced cular disease among type 2 diabetic
51. Carey PE, Halliday J, Snaar JEM, Morris b-cell proliferation in mice by inducing patients. Diabetes Care 2007;30:1212
PG, Taylor R. Direct assessment of muscle cell cycle inhibitors p16 and p18. Diabetes 1218
glycogen storage after mixed meals in 2012;61:632641 77. Suzuki A, Angulo P, Lymp J, et al. Chro-
normal and type 2 diabetic subjects. Am J 65. Ohneda M, Inman LR, Unger RH. Caloric nological development of elevated ami-
Physiol Endocrinol Metab 2003;284: restriction in obese pre-diabetic rats pre- notransferases in a nonalcoholic population.
E688E694 vents beta-cell depletion, loss of beta-cell Hepatology 2005;41:6471
52. Nobili V, Marcellini M, Devito R, et al. GLUT 2 and glucose incompetence. Dia- 78. Giannini C, Weiss R, Cali A, et al. Evidence
NAFLD in children: a prospective clinical- betologia 1995;38:173179 for early defects in insulin sensitivity and

1054 DIABETES CARE, VOLUME 36, APRIL 2013 care.diabetesjournals.org


Taylor

secretion before the onset of glucose dys- diabetic patients: long-term effects of diabetes. Eur J Public Health 1996;6:252
regulation in obese youths: a longitudinal bariatric surgery. Diabetes Care 2007;30: 256
study. Diabetes 2012;61:606614 10021004 92. Sondergaard E, Rahbek I, Srensen LP, et al.
79. Festa A, Williams K, DAgostino R Jr, 85. Adiels M, Taskinen MR, Packard C, et al. Effects of exercise on VLDL-triglyceride
Wagenknecht LE, Haffner SM. The natu- Overproduction of large VLDL particles is oxidation and turnover. Am J Physiol En-
ral course of beta-cell function in non- driven by increased liver fat content in docrinol Metab 2011;300:E939E944
diabetic and diabetic individuals: the man. Diabetologia 2006;49:755765 93. Colberg SR, Sigal RJ, Fernhall B, et al.;
Insulin Resistance Atherosclerosis Study. 86. Sjstrm L, Lindroos AK, Peltonen M, American College of Sports Medicine;
Diabetes 2006;55:11141120 et al.; Swedish Obese Subjects Study Sci- American Diabetes Association. Exercise
80. Tabak AG, Jokela M, Akbaraly TN, entic Group. Lifestyle, diabetes, and and type 2 diabetes: the American College
Brunner EJ, Kivimki M, Witte DR. Tra- cardiovascular risk factors 10 years after of Sports Medicine and the American
jectories of glycaemia, insulin sensitivity, bariatric surgery. N Engl J Med 2004;351: Diabetes Association: joint position
and insulin secretion before diagnosis 26832693 statement. Diabetes Care 2010;33:e147
of type 2 diabetes: an analysis from 87. Wilson EA, Hadden DR, Merrett JD, e167
the Whitehall II study. Lancet 2009;373: Montgomery DA, Weaver JA. Dietary man- 94. King NA, Caudwell P, Hopkins M, et al.
22152221 agement of maturity-onset diabetes. BMJ Metabolic and behavioral compensatory
81. Hanley AJ, Williams K, Festa A, et al.; 1980;280:13671369 responses to exercise interventions: barriers
Insulin Resistance Atherosclerosis Study. 88. Steven S, Lim EL, Taylor R. Population to weight loss. Obesity (Silver Spring) 2007;
Elevations in markers of liver injury and response to information on reversibility of 15:13731383
risk of type 2 diabetes: the Insulin Re- type 2 diabetes. Diabet Med. 15 January 95. Hallsworth K, Fattakhova G, Hollingsworth
sistance Atherosclerosis Study. Diabetes 2013 [Epub ahead of print] KG, et al. Resistance exercise reduces liver
2004;53:26232632 89. Delahanty LM, Pan Q, Jablonski KA, et al.; fat and its mediators in non-alcoholic fatty
82. Rudenski AS, Hadden DR, Atkinson AB, Diabetes Prevention Program Research liver disease independent of weight loss.
et al. Natural history of pancreatic islet Group. Genetic predictors of weight loss Gut 2011;60:12781283
B-cell function in type 2 diabetes mellitus and weight regain after intensive lifestyle 96. Wing RR, Phelan S. Long-term weight loss
studied over six years by homeostasis modication, metformin treatment, or maintenance. Am J Clin Nutr 2005;82
model assessment. Diabet Med 1988;5: standard care in the Diabetes Prevention (Suppl.):222S225S
3641 Program. Diabetes Care 2012;35:363 97. American Diabetes Association. Eco-
83. Matthews DR, Cull CA, Stratton IM, 366 nomic costs of diabetes in the U.S. in
Holman RR, Turner RC; UK Prospective 90. Goto Y, Nakayama Y, Yagi T. Inuence of 2007. Diabetes Care 2008;31:596615
Diabetes Study (UKPDS) Group. UKPDS the World War II food shortage on the 98. Taylor R. Pathogenesis of type 2 diabetes:
26: Sulphonylurea failure in non-insulin- incidence of diabetes mellitus in Japan. tracing the reverse route from cure to
dependent diabetic patients over six years. Diabetes 1958;7:133135 cause. Diabetologia 2008;51:17811789
Diabet Med 1998;15:297303 91. Kulenovic I, Robertson A, Grujic M, 99. Samuel VT, Shulman GI. Mechanisms for
84. Camastra S, Manco M, Mari A, et al. Beta- Suljevic E, Smajkic A. The impact of war insulin resistance: common threads and
cell function in severely obese type 2 on Sarajevans with non-insulin-dependent missing links. Cell 2012;148:852871

care.diabetesjournals.org DIABETES CARE, VOLUME 36, APRIL 2013 1055