Diabetologia (2013) 56:1489–1493

DOI 10.1007/s00125-013-2902-4

META-ANALYSIS

Progression rates from HbA1c 6.0–6.4%

and other prediabetes definitions
to type 2 diabetes: a meta-analysis
D. H. Morris & K. Khunti & F. Achana & B. Srinivasan &
L. J. Gray & M. J. Davies & D. Webb

Received: 15 February 2013 / Accepted: 15 March 2013 / Published online: 13 April 2013
# Springer-Verlag Berlin Heidelberg 2013

Abstract ADA-defined IFG (11 studies; 35.5 [26.6, 48.0]) and was
Aims/hypothesis Precise estimates of progression rates from non-significantly lower than WHO-defined IFG (34 studies;
‘prediabetes’ to type 2 diabetes are needed to optimise 47.4 [37.4, 59.8]), IGT (46 studies, 45.5 [37.8, 54.5]) and
prevention strategies for high-risk individuals. There is ac- IFG plus IGT (15 studies, 70.4 [53.8, 89.7]). Similar results
ceptance of prediabetes defined by impaired fasting glucose were seen when the data were analysed by the criteria used
(IFG) and impaired glucose tolerance (IGT), but there is to diagnose diabetes.
some controversy surrounding HbA1c-defined prediabetes Conclusions/interpretation This study provides evidence
ranges, with some favouring 6.0–6.4% (42–46 mmol/mol). that progression rates differ by prediabetes definition, which
Comparing progression rates between groups might aid this has implications for the planning and implementation of
issue, thus we aimed to accurately estimate progression rates diabetes prevention programmes. HbA1c 6.0–6.4% might
to diabetes from different prediabetes categories. identify people at a lower diabetes risk than other prediabe-
Methods Meta-analysis of prospective observational studies tes definitions, but further research is needed.
in which participants had prediabetes at baseline (ADA-
defined IFG [5.6–6.9 mmol/l], WHO-defined IFG [6.1– Keywords Diabetes mellitus, type 2 . Disease progression .
6.9 mmol/l], IGT (7.8–11.0 mmol/l) or raised HbA1c [6.0– Haemoglobin A1c protein, human . Meta-analysis .
6.4%/42–46 mmol/mol]) and were followed up for incident Prediabetic state
diabetes. Incidence rates were combined using Bayesian
random effects models. Abbreviations
Results Overall, 70 studies met the inclusion criteria. In the FBG Fasting blood glucose
six studies that used raised HbA1c, the pooled incidence rate IFG Impaired fasting glucose
(95% credible interval) of diabetes was 35.6 (15.1, 83.0) per IGT Impaired glucose tolerance
1,000 person-years. This rate was most similar to that for

Introduction
Electronic supplementary material The online version of this article
(doi:10.1007/s00125-013-2902-4) contains peer-reviewed but unedited Type 2 diabetes has adverse outcomes including early mortal-
supplementary material, which is available to authorised users.
ity [1], and is traditionally defined by the WHO on the basis of
D. H. Morris (*) : K. Khunti : B. Srinivasan : M. J. Davies : OGTTs as fasting blood glucose (FBG) ≥7.0 mmol/l or 2 h
D. Webb
glucose ≥11.1 mmol/l (glucose measured 2 h after consump-
Diabetes Research Unit, Leicester Diabetes Centre,
University of Leicester, Leicester General Hospital, tion of a standard 75 g glucose dose) [2]. Blood glucose is,
Gwendolen Road, Leicester LE5 4PW, UK however, continuous, and evidence suggests that people with
e-mail: [email protected] high, but not within diabetes range, glucose levels still risk
F. Achana : L. J. Gray
developing diabetes complications and have a high risk of
Department of Health Sciences, University of Leicester, developing overt diabetes. Therefore, this condition is often
Leicester, UK referred to as ‘prediabetes’. Controversy surrounds this
1490 Diabetologia (2013) 56:1489–1493

terminology and the use of hyperglycaemia cut-off points, but Meta-analysis The main outcome was the pooled inci-
identifying high-risk groups is useful for planning and imple- dent rate (IR) of type 2 diabetes per 1,000 person-years
menting diabetes prevention programmes. for each prediabetes definition. Data are presented as
Different prediabetes definitions exist, mostly based on pooled IR (95% credible interval [CrI]) where a CrI is
FBG (impaired fasting glucose, IFG) or 2 h glucose (im- the Bayesian equivalent of a confidence interval. The
paired glucose tolerance, IGT). The WHO 1999 criteria IRs were estimated using Bayesian random effects meta-
define prediabetes as FBG of 6.1–6.9 mmol/l (IFGWHO) analysis with vague priors in WinBUGS v1.4.3, assum-
and/or 2 h glucose of 7.8–11.0 mmol/l [2]. The ADA 2003 ing that the incident cases had a Poisson distribution
criteria define prediabetes as FBG of 5.6–6.9 mmol/l and the log IR was normally distributed. Models were
(IFGADA) [3]. There is recent interest in prediabetes defined fitted for all studies and separately by diabetes diagnos-
using HbA1c. In 2011, the WHO added HbA1c ≥6.5% tic criteria. Heterogeneity was estimated using the
(48 mmol/mol) to their diabetes definition, but conclud- between-study variance (τ2). Sources of heterogeneity
ed that there was insufficient evidence regarding an were explored using univariate meta-regression, with
HbA1c prediabetes range [4]. Conversely, the ADA have study-level covariates as the explanatory variables. As
added HbA1c of 5.7–6.4% (39–46 mmol/mol) to their an example, the WinBUGS code to fit one of the
prediabetes definition [5], while the International Expert models is in ESM Methods. The remaining WinBUGS
Committee [6] and the UK-based National Institute for code is analogous and available from the authors upon
Health and Clinical Excellence [7] support using 6.0– request.
6.4% (42–46 mmol/mol).
We conducted a meta-analysis of progression rates from
different prediabetes definitions to diabetes in observational
studies. Our aim was to provide accurate estimates of pro- Results
gression rates focusing on whether HbA1c of 6.0–6.4%
(HbA1c6.0–6.4%) had a similar progression rate to other ac- The literature search identified 70 eligible studies (ESM
cepted definitions to address the gap in knowledge regard- Fig. 1), the characteristics of which are in ESM Table 1.
ing this definition. Pooled IRs of diabetes per 1,000 person-years are given in
Table 1. When any diabetes definition was allowed, the IR
was lowest for IFGADA followed by IGT, IFGWHO and IFG+
Methods IGT. However, scrutiny of the CrIs shows that only IFG+
IGT had a significantly higher IR than the other categories,
Data sources We searched Ovid Medline (from 1993 to 10 and IFGADA, IFGWHO and IGT were not significantly dif-
February 2012) and Embase (from 1993 to Week 5 2012) ferent from each other. These results were fairly consistent
using a search strategy constructed with a clinical librarian regardless of diabetes definition.
(electronic supplementary material [ESM] Methods). It The IR for HbA1c6.0–6.4% was most similar to the IFGADA
used Medical Subject Headings and free text terms cover- rate, although not significantly different from any of the
ing ‘prediabetes’, ‘type 2 diabetes’ and ‘progression’, and other rates, when any diabetes definition was allowed and
was restricted to articles written in English and pertaining when the diabetes criteria included HbA1c. When the diabe-
to humans. Reference lists of relevant systematic reviews tes criteria included FBG, the HbA1c6.0–6.4% rate based on
and articles were hand searched. We then removed dupli- two studies was most similar to IFGWHO. No studies that
cate publications and reviewed the articles to identify included HbA1c6.0–6.4% used 2 h glucose to diagnose diabe-
studies that met the following inclusion criteria: (1) obser- tes. Heterogeneity (τ2) ranged from 0.18 to 1.10, with most
vational prospective study; (2) at least a subset of the of the prediabetes subgroups showing low to moderate
participants had IFG ADA , IFG WHO , IGT, IFG+IGT or heterogeneity (Table 1).
HbA1c6.0–6.4% at baseline; (3) the prediabetes group was The IRs were unadjusted, and therefore the effect of well-
followed up; (4) incident diabetes was reported; and (5) all established diabetes risk factors on the IRs was investigated
subjects were aged ≥18 years at baseline. Studies were through meta-regression (Table 2). The IR ratios were not
excluded if participants received bariatric surgery or any significant, except that, in the IFGWHO group, mean BMI
lifestyle or pharmacological intervention likely to affect was inversely associated with progression, and there was a
progression rates. Where a study was reported in multiple lower progression rate in Europe than in Asia in the IFGWHO
publications, the publication with the highest total person- groups (Asia: IR 60.6 [48.0, 76.2]; Europe: IR 33.0 [15.7,
years was included. Authors were contacted to obtain 66.3]) and IFG+IGT (Asia: IR 100.8 [77.9, 126.8]; Europe:
missing data. IR 53.2 [25.4, 102.8]).
Table 1 Pooled incidence rate of diabetes per 1,000 person-years according to different criteria used to diagnose prediabetes and diabetes

Prediabetes definition Criteria used to diagnose diabetes

Any Included FBG Included 2 h glucose Included HbA1c

N IR (95% CrI)a τ2 N IR (95% CrI) τ2 N IR (95% CrI) τ2 N IR (95% CrI) τ2

IFGADAb 11 35.54 (26.60, 48.02) 0.22 11 35.54 (26.60, 48.02) 0.22 4 33.04 (11.43, 102.10) 1.42 2 37.18 (0.22, 6,399) 10.3
IFGWHOc 34 47.40 (37.39, 59.81) 0.42 31 48.56 (39.56, 59.11) 0.26 17 41.58 (28.24, 60.88) 0.53 1 66.4
IGT 46 45.46 (37.76, 54.52) 0.35 37 45.71 (37.68, 55.11) 0.31 39 47.39 (39.63, 56.53) 0.28 1 96.68
Diabetologia (2013) 56:1489–1493

IFG and IGT 15 70.36 (53.78, 89.71) 0.18 14 72.91 (55.26, 93.62) 0.18 12 74.92 (55.45, 98.43) 0.20 0
Raised HbA1cd 6 35.62 (15.14, 82.97) 1.10 2 56.55 (0.05, 62,040) 19.28 0 4 25.63 (10.06, 77.56) 1.21
a
IR shows the pooled incidence rate of diabetes per 1,000 person-years
b
IFG defined as fasting glucose 5.6–6.9 mmol/l
c
IFG defined as fasting glucose 6.1–6.9 mmol/l
d
HbA1c in the range 6.0–6.4% (42–46 mmol/mol)
1491
1492 Diabetologia (2013) 56:1489–1493

Table 2 IR ratios (95% credible interval) showing the effect of study-level covariates on IRs of diabetes

Study-level covariate Prediabetes group

IFGADA IFGWHO IGT IFG and IGT Raised HbA1ca

Mean age, years 1.00 (0.94, 1.06) 0.97 (0.94, 1.00) 0.99 (0.97, 1.01) 0.99 (0.95, 1.03) 0.95 (0.75, 1.21)
Year published, years 0.95 (0.82, 1.10) 1.02 (0.94, 1.09) 0.99 (0.96, 1.03) 1.04 (0.96, 1.12) 0.73 (0.32, 1.77)
Male, % 1.00 (0.98, 1.01) 1.01 (1.00, 1.02) 0.99 (0.99, 1.00) 1.01 (0.99, 1.03) 1.02 (0.99, 1.06)
White European, % 1.00 (0.96, 1.03) 1.00 (0.99, 1.00) 0.99 (0.99, 1.00) 1.00 (0.98, 1.01) 0.99 (0.92, 1.06)
Family history of diabetes, % 1.01 (0.99, 1.02) 1.00 (0.97, 1.03) 1.01 (1.00, 1.01) 0.99 (0.97, 1.01) 0.95 (0.82, 1.09)
Mean BMI, kg/m2 1.04 (0.93, 1.17) 0.86 (0.75, 0.99) 1.03 (0.92, 1.16) 0.95 (0.77, 1.17) 0.89 (0.46, 1.77)
Continent
Asia Reference Reference Reference Reference Reference
Europe 0.42 (0.17, 1.03) 0.57 (0.33, 0.96) 0.65 (0.41, 1.03) 0.55 (0.35, 0.84) 0.21 (0.02, 2.03)
Americas 1.04 (0.54, 2.02) 0.76 (0.36, 1.59) 1.01 (0.63, 1.61) 0.62 (0.37, 1.00) 0.30 (0.02, 4.96)
Other No studies 0.69 (0.31, 1.55) 1.18 (0.64, 2.19) No studies 0.59 (0.03, 9.41)
a
6.0–6.4% (37–46 mmol/mol)

Discussion HbA1c6.0–6.4% is associated with a slightly lower diabetes risk

than IFGWHO and IGT and reinforces the need for further
Accurately describing progression rates to type 2 diabetes is research establishing the predictive capacity of HbA1c6.0–6.4%.
clinically important for accurately identifying people at par- Within some subgroups, there was high heterogeneity,
ticularly high risk and for effectively planning interventions which we explored using study-level covariates. Few cova-
and monitoring of these people. Thus, we used a meta- riates were significant, suggesting that the heterogeneity
analysis to pool existing estimates of progression rates with- was either due to covariates that were not considered or
in different prediabetes definitions. chance. The way in which diabetes was defined appeared
The groups identified by IFG and IGT have relatively to play a role because on the whole heterogeneity was
little overlap, leading to suggestions that fasting and post- reduced when analyses were stratified by diabetes criteria.
challenge hyperglycaemia are driven by different biological Glucose and HbA1c appear to detect different diabetes pop-
mechanisms [8]. It is suggested that IGT is the result of ulations [8]. The way in which prediabetes and diabetes
excessive endogenous glucose production (insulin action were defined were closely related, which resulted in very
defect) in combination with beta cell dysfunction (insulin small numbers of studies in some subgroups, making com-
secretion defect). Conversely, after food ingestion, this parisons difficult. However, progression rates were similar
mainly hepatic glucose output is appropriately suppressed when fasting glucose or 2 h glucose was included. Too few
in individuals with isolated IFG, suggesting that an insulin studies used HbA1c as a diabetes criterion to make useful
secretory defect is mainly responsible for resultant plasma conclusions, suggesting that this is an area where future
hyperglycaemia in this condition. Our findings regarding research is required, particularly when it is considered that
IRs in IFG and IGT groups are consistent with a systematic HbA1c is now commonly being used to diagnose diabetes in
review conducted in 2007 [9], with the ADA’s finding that clinical practice.
their IFG definition had a lower specificity but a higher Our conclusions have limitations. The analysis was con-
sensitivity than the WHO’s [10], and with the knowledge fined to existing prediabetes categories, did not include
that having IFG+IGT is typically associated with later progression to diabetes from normoglycaemia or regression
phases of glucose intolerance [8]. to normoglycaemia from prediabetes, and heterogeneity
Since HbA1c was added to the diabetes criteria [4], there probably contributed to some overlap of CrIs, thereby influ-
has been interest in prediabetes defined using HbA1c, but encing outcome comparisons. Furthermore, the diabetes
current guidelines conflict about which definition to use, definition in some studies was not restricted to biochemical
primarily because of insufficient data [5–7]. Some expert confirmation, but could utilise physician diagnosis or med-
committees favour the range 6.0–6.4% [6, 7]. In that group, ication initiation as diabetes end points. The validity of
we estimated that 36 new diagnoses of diabetes would be directly comparing populations whose primary outcome is
expected per 1,000 person-years. Although the wide CrIs defined by different criteria is justified here because the aim
make interpretation difficult, it appears that this measure of the study was to report cumulative progression rates
relates most closely to the IFGADA category. Our data suggest using established and internationally accepted criteria.
Diabetologia (2013) 56:1489–1493 1493

While acknowledging the inherent difficulties associated declare that there is no duality of interest associated with their contri-
bution to this manuscript.
with combining population-level observational datasets in
meta-analyses of this kind, we feel this study has major Contribution statement KK, MJD and BS conceived the study.
strengths. To our knowledge, this is the largest predictive DHM, FA and BS designed the literature search. DHM, FA, DW and
meta-analysis of IFG, IGT and raised HbA1c categories. LJG conducted the literature search. DHM and FA extracted the data.
DHM performed the data analysis with contributions from LG and FA.
Moreover, rigorous methodology ensured literature searches
DHM and DW wrote the first draft and all authors were involved in the
and statistical analyses met accepted standards for meta- data interpretation, writing and revising the manuscript and approved
analyses. the final version. DHM had full access to all the data in the study and
We provided pooled estimates of progression rates from takes responsibility for the integrity of the data and the accuracy of the
data analysis.
prediabetes to diabetes that suggest that rates were lowest
for IFG, slightly higher for IGT, and highest for IFG+IGT.
HbA1c6.0–6.4% had a similar diabetes risk to IFGADA. Further
References
investigation is justified since non-significance meant that
our results were not conclusive. These findings suggest that
different management strategies might be required in future 1. Massi-Benedetti M (2002) The cost of diabetes type II in Europe.
The CODE-2 study: preface. Diabetologia 45:S1–S4
prevention programmes. 2. World Health Organization (1999) Definition, diagnosis and clas-
sification of diabetes mellitus and its complications. Part 1: diag-
Acknowledgements The authors would like to thank Sarah Sutton nosis and classification of diabetes mellitus. World Health
(University Hospitals of Leicester Libraries, UK) for her valuable input Organization, Geneva
into designing the literature search, Tom Leigh (Leicester Diabetes 3. American Diabetes Association (2004) Diagnosis and classifica-
Centre, University Hospitals of Leicester) for his help with performing tion of diabetes mellitus. Diabetes Care 27:s5–s10
the literature review, Beverley Balkau, Richard Legro, Inna Misnikova 4. World Health Organization (2011) Use of glycated haemoglobin
and Soraya Soulimane for providing data that were not available in the (HbA1c) in the diagnosis of diabetes mellitus. World Health Or-
original study reports, and the following authors for providing data for ganization, Geneva
an earlier version of this manuscript: Simona Bo, Hertzel Gerstein, 5. American Diabetes Association (2010) Diagnosis and classifica-
Rury Holman on behalf of the NAVIGATOR Executive Committee, tion of diabetes mellitus. Diabetes Care 33:S62–S69
Yonghong Huan, Allen Kunselman, Pia Pajunen and Markku Peltonen. 6. The International Expert Committee (2009) International Expert
Committee report on the role of the A1C assay in the diagnosis of
Funding This work was supported by funding from the Department diabetes. Diabetes Care 32:1327–1334
of Health, the University Hospitals of Leicester Diabetes Research 7. National Institute for Health and Clinical Excellence (2012)
Fund and Diabetes UK, and was conducted as part of activities sup- Public health draft guidance. Preventing type 2 diabetes: risk
ported by the National Institute for Health Research Collaboration for identification and interventions for individuals at high risk.
Leadership in Applied Health Research and Care–Leicestershire, Available from http://guidance.nice.org.uk/PHG/Wave19/62.
Northamptonshire and Rutland (NIHR CLAHRC–LNR) and The Accessed 21 March 2013
NIHR Leicester–Loughborough Diet, Lifestyle and Physical Activity 8. Mostafa SA, Davies MJ, Srinivasan BT, Carey ME, Webb D,
Biomedical Research Unit, which is a partnership between University Khunti K (2010) Should glycated haemoglobin (HbA1c) be used
Hospitals of Leicester NHS Trust, Loughborough University and the to detect people with type 2 diabetes mellitus and impaired glucose
University of Leicester. The study funders had no role in the collection, regulation? Postgrad Med J 86:656–662
management, analysis or interpretation of the data, or in the prepara- 9. Gerstein HC, Santaguida P, Raina P et al (2007) Annual
tion, review, or approval of the manuscript. incidence and relative risk of diabetes in people with various
categories of dysglycemia: a systematic overview and meta-
analysis of prospective studies. Diabetes Res Clin Pract
Duality of interest K. Khunti and M.J. Davies are members of the 78:305–312
National Institute for Health and Clinical Excellence (NICE) commit- 10. The Expert Committee on the Diagnosis and Classification of
tee and are also part of the Department of Health group recommending Diabetes Mellitus (2003) Follow-up report on the diagnosis of
HbA1c for the identification of high-risk people. All other authors diabetes mellitus. Diabetes Care 26:3160–3167