It is an unexpected but pleasant surprise when new clinical relationships are identified, and one of the most interesting is the inverse association between LDL cholesterol (LDLc) and type 2 diabetes (T2D) risk. Evidence from both randomized clinical trials and genetic studies indicates that regulation of plasma lipids and glycemic control is more closely linked than previously assumed, yet in a counterintuitive, one could even say paradoxical, manner. Meta-analyses of randomized clinical trials have found that drugs designed to reduce LDLc, in addition to their hypolipidemic and cardioprotective effects, appear to also modestly increase T2D risk (1,2). Furthermore, naturally occurring genetic variation in molecular targets of LDLc-lowering therapy, such as genetic variants in or near HMGCR, NCP1L1, and PCSK9 genes, have been found to be associated with impaired insulin sensitivity and new-onset T2D, particularly among people with impaired fasting glucose levels (36). Further supporting that this is a fundamental biologic relationship, individuals with familial hypercholesterolemia, a dominantly inherited disease characterized by high plasma levels of LDLc due to genetic mutations in LDLR or APOB genes, appear to have a lower prevalence of diabetes than unaffected relatives (7). However, not all genetic variants that raise LDLc have similar effects on glycemic control (8). This suggests that the mechanism by which LDLc is reduced might have relevant implications for glycemic deterioration and reveal potential important mechanisms for diabetogenesis in general.

As reported in this issue of Diabetes, Klimentidis et al. (9) conducted a study to examine the phenotypic and genotypic relationships between LDLc and T2D (Fig. 1). Using data from the UK Biobank (n = 431,167), they confirmed findings from previous reports that LDLc is inversely associated with T2D prevalence, in this case with an odds ratio of 0.41 [95% CI 0.39, 0.43] per each mmol/L increase in LDLc, which they term an “opposite direction of effect.” Although the magnitude of their observed association was higher than in other studies considering T2D incidence as opposed to T2D prevalence (1,2,10,11), these findings remained similar in several sensitivity analyses taking into account potential bias such as incomplete case ascertainment or the presence of a collider. The phenotypic associations also showed a paradoxical relationship of increased LDLc with increased HbA1c, which may be an indication of the difficulties of studying the relationship between these complex phenotypes in a cross-sectional study.

Figure 1

Overview of main study findings. In a cross-sectional study within the UK Biobank (n = 431,167), Klimentidis et al. (9) reported that LDLc is inversely associated with T2D prevalence (odds ratio 0.41 [95% CI 0.39, 0.43] per each mmol/L increase in LDLc). Using genetic data from UK Biobank and DIAGRAM (n = 898,130), Klimentidis et al. identified 44 genomic regions, with opposite associations between LDLc and T2D (31 of them then replicated) enriched for genes associated with NAFLD. Their findings suggest that the diabetogenic effect of lipid-lowering medications is in part mediated by increased liver fat content.

Figure 1

Overview of main study findings. In a cross-sectional study within the UK Biobank (n = 431,167), Klimentidis et al. (9) reported that LDLc is inversely associated with T2D prevalence (odds ratio 0.41 [95% CI 0.39, 0.43] per each mmol/L increase in LDLc). Using genetic data from UK Biobank and DIAGRAM (n = 898,130), Klimentidis et al. identified 44 genomic regions, with opposite associations between LDLc and T2D (31 of them then replicated) enriched for genes associated with NAFLD. Their findings suggest that the diabetogenic effect of lipid-lowering medications is in part mediated by increased liver fat content.

To identify genetic variants with opposite effects on LDLc and T2D prevalence, Klimentidis et al. conducted a cleverly designed two-stage genome-wide association study. They first identified genetic variants associated with LDLc in UK Biobank. Then, to confirm those dual LDLc-T2D variant associations, they used T2D summary statistics data from the Diabetes Genetics Replication and Meta-Analysis (DIAGRAM) consortium (n = 898,130). This led to the initial identification of 44 genomic regions with opposite associations between LDLc and T2D, and 31 of them then replicated in the independent data sets for their association with LDLc. A number of the genomic regions identified by Klimentidis et al. were previously known or suspected to be inversely associated with circulating LDLc and T2D (HMGCR, NPC1L1, APOE), but the authors also identified 14 genomic regions without evidence of previous association with LDLc or T2D. Thus, first of all, this is a novel way to identify new LDLc and T2D loci. Computational characterization of identified genomic regions suggests that genetic variants that have an opposite effect on LDLc and T2D are enriched for genes associated with nonalcoholic fatty liver disease (NAFLD). Of particular interest is the observation that for some of the identified genomic regions, the same exact variant that has the joint effect on LDLc and T2D is the variant associated with increased NAFLD (i.e., GCKR, PNPLA3, PPP1R3B, or TM6SF2), highlighting the relevance of liver metabolism on plasma lipids and glycemic control.

There are some limitations to the study. First, this version of UK Biobank is a cross-sectional study, and it will be of interest to repeat these analyses when prospective data become available. For example, it is possible that T2D cases are more likely to be newly diagnosed patients not yet under lipid-lowering therapy, those with intolerance to lipid-lowering medications, people misreporting lipid-lowering medications, or older T2D individuals not requiring lipid-lowering medications. Propensity score analyses were implemented to account for this potential bias, but confounding could still exist (12). Second, this is not a traditional joint-phenotype genetic study in which the same participants have the phenotype of interest. By including data for LDLc from participants within the UK Biobank, and a separate data set to investigate whether these genetic variants associated with LDLc have a divergent effect on T2D, it is possible that differences in genome-wide association study characteristics may introduce some noise. Third, while mapping variants to genes is difficult, and only some of these loci exhibited colocalization of the association signals, it is still the case that we often infer that the closest gene is the most likely causal gene. Emerging data indicate that this is not always the case (13).

Overall, findings from Klimentidis et al. provide a new perspective on the debate regarding the intriguing physiological relationship between lipids and dysglycemia and put liver metabolism in the spotlight. Animal and human physiological studies have found that fat accumulation in the liver leads to hepatic insulin resistance and that direct and indirect mechanisms exist to control insulin’s regulation of hepatic glucose and fat production (14,15). The observation that T2D signals identified for their dual association with low LDLc are mainly insulin-resistance loci, as opposite to the growing number of T2D loci primarily associated with insulin secretion, is well aligned with previous physiological data linking insulin resistance with hepatic fatty acid uptake and adipose tissue dysfunction. In this context, lipid-lowering strategies promoting adipose tissue expandability might have relevant implications to reduce glycemic deterioration associated with reducing LDLc, as it has been recently demonstrated for both gain- and loss-of-function variants in the LPL gene and T2D risk (16). Evidence from the current study may foster new lines of investigation to gain insights, not only into the underlying mechanisms responsible for the diabetogenic effect of LDLc-lowering medications, but into the etiology of T2D itself. Such knowledge will hopefully be used to inform public health and individual strategies to leverage more personalized and efficient approaches to manage dyslipidemia among people with impaired fasting glucose levels.

See accompanying article, p. 2194.

Funding. J.M. is supported by European Commission Horizon 2020 program grant H2020-MSCA-IF-2015-703787 and National Institutes of Health (NIH) grant P30 DK040561. J.I.R. is supported in part by NIH National Center for Advancing Translational Sciences (NCATS) UCLA CTSI Grant Number UL1TR001881 and National Institute of Diabetes and Digestive and Kidney Diseases Diabetes Research Center grant DK063491 to the Southern California Diabetes Endocrinology Research Center. J.I.R. is also supported in part by NIH grant R01-HL151855–01.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

1.
Sattar
N
,
Preiss
D
,
Murray
HM
, et al
.
Statins and risk of incident diabetes: a collaborative meta-analysis of randomised statin trials
.
Lancet
2010
;
375
:
735
742
2.
Swerdlow
DI
,
Preiss
D
,
Kuchenbaecker
KB
, et al.;
DIAGRAM Consortium
;
MAGIC Consortium
;
InterAct Consortium
.
HMG-coenzyme A reductase inhibition, type 2 diabetes, and bodyweight: evidence from genetic analysis and randomised trials
.
Lancet
2015
;
385
:
351
361
3.
Lotta
LA
,
Sharp
SJ
,
Burgess
S
, et al
.
Association between low-density lipoprotein cholesterol-lowering genetic variants and risk of type 2 diabetes: a meta-analysis
.
JAMA
2016
;
316
:
1383
1391
4.
Ference
BA
,
Robinson
JG
,
Brook
RD
, et al
.
Variation in PCSK9 and HMGCR and risk of cardiovascular disease and diabetes
.
N Engl J Med
2016
;
375
:
2144
2153
5.
Schmidt
AF
,
Swerdlow
DI
,
Holmes
MV
, et al.;
LifeLines Cohort study group
;
UCLEB consortium
.
PCSK9 genetic variants and risk of type 2 diabetes: a mendelian randomisation study
.
Lancet Diabetes Endocrinol
2017
;
5
:
97
105
6.
Goodarzi
MO
,
Rotter
JI
.
Genetics insights in the relationship between type 2 diabetes and coronary heart disease
.
Circ Res
2020
;
126
:
1526
1548
7.
Besseling
J
,
Kastelein
JJP
,
Defesche
JC
,
Hutten
BA
,
Hovingh
GK
.
Association between familial hypercholesterolemia and prevalence of type 2 diabetes mellitus
.
JAMA
2015
;
313
:
1029
1036
8.
Liu
DJ
,
Peloso
GM
,
Yu
H
, et al.;
Charge Diabetes Working Group
;
EPIC-InterAct Consortium
;
EPIC-CVD Consortium
;
GOLD Consortium
;
VA Million Veteran Program
.
Exome-wide association study of plasma lipids in >300,000 individuals
.
Nat Genet
2017
;
49
:
1758
1766
9.
Klimentidis
YC
,
Arora
A
,
Newell
M
, et al
.
Phenotypic and genetic characterization of lower LDL cholesterol and increased type 2 diabetes risk in the UK Biobank
.
Diabetes
2020
;
69
:
2194
2205
10.
Andersson
C
,
Lyass
A
,
Larson
MG
,
Robins
SJ
,
Vasan
RS
.
Low-density-lipoprotein cholesterol concentrations and risk of incident diabetes: epidemiological and genetic insights from the Framingham Heart Study
.
Diabetologia
2015
;
58
:
2774
2780
11.
Feng
Q
,
Wei
W-Q
,
Chung
CP
, et al
.
Relationship between very low low-density lipoprotein cholesterol concentrations not due to statin therapy and risk of type 2 diabetes: a US-based cross-sectional observational study using electronic health records
.
PLoS Med
2018
;
15
:
e1002642
12.
Munafò
MR
,
Tilling
K
,
Taylor
AE
,
Evans
DM
,
Davey Smith
G
.
Collider scope: when selection bias can substantially influence observed associations
.
Int J Epidemiol
2018
;
47
:
226
235
13.
Barroso
I
,
McCarthy
MI
.
The genetic basis of metabolic disease
.
Cell
2019
;
177
:
146
161
14.
Birkenfeld
AL
,
Shulman
GI
.
Nonalcoholic fatty liver disease, hepatic insulin resistance, and type 2 diabetes
.
Hepatology
2014
;
59
:
713
723
15.
Stefan
N
,
Häring
H-U
,
Cusi
K
.
Non-alcoholic fatty liver disease: causes, diagnosis, cardiometabolic consequences, and treatment strategies
.
Lancet Diabetes Endocrinol
2019
;
7
:
313
324
16.
Stitziel
NO
,
Stirrups
KE
,
Masca
NGD
, et al.;
Myocardial Infarction Genetics and CARDIoGRAM Exome Consortia Investigators
.
Coding variation in ANGPTL4, LPL, and SVEP1 and the risk of coronary disease
.
N Engl J Med
2016
;
374
:
1134
1144
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/content/license.