OBJECTIVE

The increased risk of type 2 diabetes in women with glucose intolerance in pregnancy is mediated by deterioration of their β-cell function, which occurs as early as the first year postpartum. We thus sought to identify early determinants of their declining β-cell function.

RESEARCH DESIGN AND METHODS

Women with recent gestational glucose intolerance (166) underwent oral glucose tolerance test at 3 and 12 months postpartum. They were stratified into those in whom β-cell function (Insulin Secretion-Sensitivity Index-2 [ISSI-2]) declined over this time (decliners; n = 92) and those in whom it did not (nondecliners; n = 74).

RESULTS

Between 3 and 12 months, hepatic insulin sensitivity (1/homeostasis model assessment of insulin resistance [HOMA-IR]) decreased in decliners but not in nondecliners. Over this time, the change in 1/HOMA-IR emerged as an independent predictor of the change in ISSI-2 (t = 5.5; P < 0.0001). Increased hepatic insulin sensitivity independently predicted a lower likelihood of declining β-cell function (odds ratio = 0.13 [95% CI 0.06–0.29]; P < 0.0001).

CONCLUSIONS

Hepatic insulin resistance is an early determinant of declining β-cell function after gestational dysglycemia.

Women who develop gestational diabetes mellitus (GDM) and gestational impaired glucose tolerance (GIGT) have a chronic defect in β-cell function that underlies their presentation with dysglycemia in response to the severe insulin resistance of late pregnancy (1,2). After the pregnancy, their increased risk of developing type 2 diabetes is mediated by progressive deterioration of the β-cell function over time, which occurs as early as the first year postpartum (35). Thus, we sought to identify early determinants of declining β-cell function in a cohort of women with GDM and GIGT evaluated at 3 and 12 months postpartum.

This analysis was conducted within an ongoing observational cohort study in which women recruited at the time of antepartum GDM screening were undergoing longitudinal evaluation in pregnancy and in the postpartum years. The study protocol has been described previously (2,5,6). In brief, participants undergo an oral glucose tolerance test (OGTT) in late pregnancy, enabling classification of gestational glucose tolerance (GDM, GIGT, normal). At 3 and 12 months postpartum, they undergo reassessment, including 2-h 75-g OGTT. The study has been approved by the institutional research ethics board, and all participants have provided written informed consent. As reported previously (5), among the first 392 women who completed their 12-month visit, there were 182 women with either GDM or GIGT. The current analysis was restricted to 166 women with GDM or GIGT, after exclusion of those who had diabetes at 3 months (n = 5) or any missing glucose/insulin values on OGTTs (n = 11) (Supplementary Fig. 1).

As described previously (5), β-cell function was assessed with the Insulin Secretion-Sensitivity Index-2 (ISSI-2) and whole-body insulin sensitivity was measured by Matsuda index. The homeostasis model assessment of insulin resistance (HOMA-IR) primarily reflects hepatic insulin resistance (79); its reciprocal (1/HOMA-IR) provides a measure of hepatic insulin sensitivity. As a secondary measure of hepatic insulin sensitivity, fasting C-peptide:insulin was calculated. Fasting C-peptide:insulin is a measure of steady-state hepatic insulin extraction, which decreases with worsening hepatic insulin sensitivity (1013).

Statistical analyses were performed using SAS 9.2 (SAS Institute, Cary, NC). Univariate differences were assessed between women in whom β-cell function declined between 3 and 12 months and those in whom it did not (Table 1 and Supplementary Table 1). Univariate correlations between the baseline-adjusted change in ISSI-2 and potential predictors thereof were assessed by Spearman analysis (Supplementary Table 2). Multiple linear regression analyses were performed to identify predictors of the change in ISSI-2 between 3 and 12 months postpartum (Supplementary Table 3). Logistic regression analysis was performed to identify independent predictors of declining ISSI-2 (Supplementary Table 4).

Table 1

Demographic, clinical, and metabolic characteristics of study population stratified into two groups: women in whom β-cell function did not decline between 3 and 12 months postpartum (nondecliners) and women in whom β-cell function declined between 3 and 12 months postpartum (decliners)

NondeclinersDeclinersP
n 74 92  
At 3 months postpartum    
 Age (years) 35.3 [31.8–38.5] 35.2 [33.8–38.3] 0.4204 
 Ethnicity (%)   0.4097 
  White 77.0 72.8  
  Asian 13.5 10.9  
  Other 9.5 16.3  
 Family history of diabetes (%) 62.2 53.3 0.2492 
 Current smoking (%) 4.1 6.5 0.7325 
 Current breastfeeding (%) 97.3 91.3 0.1876 
 Blood pressure (mmHg)    
  Systolic 108 [104–117] 111 [103–116] 0.6593 
  Diastolic 65 [59–70] 64 [59–71] 0.9634 
 Physical activity    
  Sport index 2.0 [1.8–2.5] 1.8 [1.5–2.5] 0.6019 
  Leisure-time index 2.8 [2.5–3.3] 2.8 [2.5–3.3] 0.8009 
 Waist circumference (cm) 87 [82–99] 90 [84–96] 0.5303 
 Weight (kg) 70.2 [59.0–81.0] 69.9 [62.1–80.7] 0.7014 
 BMI (kg/m226.5 [23.2–31.2] 26.6 [24.0–30.2] 0.6745 
 Fasting insulin (pmol/L) 36.5 [28.0–59.0] 26.0 [20.0–49.5] 0.0130 
 Whole-body insulin sensitivity    
  Matsuda index 8.3 [5.7–14.0] 9.8 [6.0–12.9] 0.5413 
 Hepatic insulin sensitivity    
  1/HOMA-IR 0.94 [0.56–1.32] 1.3 [0.65–1.71] 0.0112 
  Fasting C-peptide:insulin 15.6 [12.9–21.5] 18.5 [14.5–27.0] 0.0115 
 β-Cell function    
  ISSI-2 785 [624–952] 914 [748–1,207] 0.0004 
 OGTT    
  Fasting glucose (mmol/L) 4.8 [4.3–5.0] 4.7 [4.3–5.0] 0.1234 
  30-min glucose (mmol/L) 8.4 [7.4–9.5] 8.7 [7.5–9.8] 0.5027 
  1-h glucose (mmol/L) 8.6 [7.1–10.1] 8.7 [7.2–10.0] 0.9948 
  2-h glucose (mmol/L) 6.4 [5.3–7.3] 6.7 [5.5–7.9] 0.4479 
  AUCgluc 14.1 [12.3–15.8] 14.2 [12.4–15.9] 0.9288 
 Glucose tolerance status (%)   0.1356 
  NGT 78.4 72.8  
  Isolated IFG 2.7  
  Isolated IGT 17.6 27.2  
  Combined IFG and IGT 1.4  
At 12 months postpartum    
 Current breastfeeding (%) 96.0 83.7 0.0123 
 Physical activity    
  Sport index 2.1 [1.8–3.0] 2.0 [1.8–2.5] 0.3252 
  Leisure-time index 3.0 [2.8–3.5] 3.0 [2.8–3.3] 0.7867 
  Work index 2.9 [2.5–3.4] 2.9 [2.4–3.4] 0.7499 
 Waist circumference (cm) 85 [79–96] 89 [82–95] 0.2181 
 Weight (kg) 68.1 [56.5–78.0] 68.3 [59.3–78.7] 0.3084 
 BMI (kg/m225.4 [21.5–29.4] 25.9 [23.4–29.3] 0.2976 
 Fasting insulin (pmol/L) 31.0 [20.0–48.0] 45.5 [27.5–73.0] 0.0015 
 Whole-body insulin sensitivity    
  Matsuda index 9.1 [5.8–13.9] 7.2 [4.4–10.9] 0.0436 
 Hepatic insulin sensitivity    
  1/HOMA-IR 1.11 [0.73–1.68] 0.73 [0.45–1.18] 0.0015 
  Fasting C-peptide:insulin 17.7 [14.0–23.9] 15.1 [10.9–20.9] 0.0116 
 β-Cell function    
  ISSI-2 1,034 [825–1,241] 710 [572–902] <0.0001 
 OGTT    
  Fasting glucose (mmol/L) 4.8 [4.5–4.9] 4.9 [4.6–5.2] 0.0570 
  30-min glucose (mmol/L) 8.0 [7.1–9.3] 8.7 [7.8–9.8] 0.0070 
  1-h glucose (mmol/L) 7.4 [5.8–9.2] 9.1 [7.4–10.5] <0.0001 
  2-h glucose (mmol/L) 5.9 [4.9–7.3] 6.7 [5.8–7.9] 0.0007 
  AUCgluc 12.6 [10.8–14.8] 14.8 [12.7–16.8] <0.0001 
 Glucose tolerance status (%)   0.3613 
  NGT 78.4 73.9  
  Isolated IFG 1.4  
  Isolated IGT 20.3 23.9  
  Combined IFG and IGT  
  Diabetes 2.2  
Baseline-adjusted changes between 3 and 12 months    
 Change in sport index 0.16 [0.07] 0.26 [0.08] 0.3903 
 Change in leisure-time index 0.16 [0.05] 0.24 [0.06] 0.2940 
 Change in waist −1.3 [0.7] −2.1 [0.8] 0.4594 
 Change in BMI −0.5 [0.2] −1.1 [0.2] 0.0585 
 Change in weight −1.5 [0.5] −2.9 [0.5] 0.0482 
 Change in Matsuda index −0.8 [0.6] 0.2 [0.7] 0.2701 
 Change in 1/HOMA-IR −0.31 [0.05] 0.07 [0.05] <0.0001 
 Change in fasting C-peptide:insulin −2.6 [1.1] 1.6 [1.2] 0.0107 
NondeclinersDeclinersP
n 74 92  
At 3 months postpartum    
 Age (years) 35.3 [31.8–38.5] 35.2 [33.8–38.3] 0.4204 
 Ethnicity (%)   0.4097 
  White 77.0 72.8  
  Asian 13.5 10.9  
  Other 9.5 16.3  
 Family history of diabetes (%) 62.2 53.3 0.2492 
 Current smoking (%) 4.1 6.5 0.7325 
 Current breastfeeding (%) 97.3 91.3 0.1876 
 Blood pressure (mmHg)    
  Systolic 108 [104–117] 111 [103–116] 0.6593 
  Diastolic 65 [59–70] 64 [59–71] 0.9634 
 Physical activity    
  Sport index 2.0 [1.8–2.5] 1.8 [1.5–2.5] 0.6019 
  Leisure-time index 2.8 [2.5–3.3] 2.8 [2.5–3.3] 0.8009 
 Waist circumference (cm) 87 [82–99] 90 [84–96] 0.5303 
 Weight (kg) 70.2 [59.0–81.0] 69.9 [62.1–80.7] 0.7014 
 BMI (kg/m226.5 [23.2–31.2] 26.6 [24.0–30.2] 0.6745 
 Fasting insulin (pmol/L) 36.5 [28.0–59.0] 26.0 [20.0–49.5] 0.0130 
 Whole-body insulin sensitivity    
  Matsuda index 8.3 [5.7–14.0] 9.8 [6.0–12.9] 0.5413 
 Hepatic insulin sensitivity    
  1/HOMA-IR 0.94 [0.56–1.32] 1.3 [0.65–1.71] 0.0112 
  Fasting C-peptide:insulin 15.6 [12.9–21.5] 18.5 [14.5–27.0] 0.0115 
 β-Cell function    
  ISSI-2 785 [624–952] 914 [748–1,207] 0.0004 
 OGTT    
  Fasting glucose (mmol/L) 4.8 [4.3–5.0] 4.7 [4.3–5.0] 0.1234 
  30-min glucose (mmol/L) 8.4 [7.4–9.5] 8.7 [7.5–9.8] 0.5027 
  1-h glucose (mmol/L) 8.6 [7.1–10.1] 8.7 [7.2–10.0] 0.9948 
  2-h glucose (mmol/L) 6.4 [5.3–7.3] 6.7 [5.5–7.9] 0.4479 
  AUCgluc 14.1 [12.3–15.8] 14.2 [12.4–15.9] 0.9288 
 Glucose tolerance status (%)   0.1356 
  NGT 78.4 72.8  
  Isolated IFG 2.7  
  Isolated IGT 17.6 27.2  
  Combined IFG and IGT 1.4  
At 12 months postpartum    
 Current breastfeeding (%) 96.0 83.7 0.0123 
 Physical activity    
  Sport index 2.1 [1.8–3.0] 2.0 [1.8–2.5] 0.3252 
  Leisure-time index 3.0 [2.8–3.5] 3.0 [2.8–3.3] 0.7867 
  Work index 2.9 [2.5–3.4] 2.9 [2.4–3.4] 0.7499 
 Waist circumference (cm) 85 [79–96] 89 [82–95] 0.2181 
 Weight (kg) 68.1 [56.5–78.0] 68.3 [59.3–78.7] 0.3084 
 BMI (kg/m225.4 [21.5–29.4] 25.9 [23.4–29.3] 0.2976 
 Fasting insulin (pmol/L) 31.0 [20.0–48.0] 45.5 [27.5–73.0] 0.0015 
 Whole-body insulin sensitivity    
  Matsuda index 9.1 [5.8–13.9] 7.2 [4.4–10.9] 0.0436 
 Hepatic insulin sensitivity    
  1/HOMA-IR 1.11 [0.73–1.68] 0.73 [0.45–1.18] 0.0015 
  Fasting C-peptide:insulin 17.7 [14.0–23.9] 15.1 [10.9–20.9] 0.0116 
 β-Cell function    
  ISSI-2 1,034 [825–1,241] 710 [572–902] <0.0001 
 OGTT    
  Fasting glucose (mmol/L) 4.8 [4.5–4.9] 4.9 [4.6–5.2] 0.0570 
  30-min glucose (mmol/L) 8.0 [7.1–9.3] 8.7 [7.8–9.8] 0.0070 
  1-h glucose (mmol/L) 7.4 [5.8–9.2] 9.1 [7.4–10.5] <0.0001 
  2-h glucose (mmol/L) 5.9 [4.9–7.3] 6.7 [5.8–7.9] 0.0007 
  AUCgluc 12.6 [10.8–14.8] 14.8 [12.7–16.8] <0.0001 
 Glucose tolerance status (%)   0.3613 
  NGT 78.4 73.9  
  Isolated IFG 1.4  
  Isolated IGT 20.3 23.9  
  Combined IFG and IGT  
  Diabetes 2.2  
Baseline-adjusted changes between 3 and 12 months    
 Change in sport index 0.16 [0.07] 0.26 [0.08] 0.3903 
 Change in leisure-time index 0.16 [0.05] 0.24 [0.06] 0.2940 
 Change in waist −1.3 [0.7] −2.1 [0.8] 0.4594 
 Change in BMI −0.5 [0.2] −1.1 [0.2] 0.0585 
 Change in weight −1.5 [0.5] −2.9 [0.5] 0.0482 
 Change in Matsuda index −0.8 [0.6] 0.2 [0.7] 0.2701 
 Change in 1/HOMA-IR −0.31 [0.05] 0.07 [0.05] <0.0001 
 Change in fasting C-peptide:insulin −2.6 [1.1] 1.6 [1.2] 0.0107 

Baseline-adjusted changes are shown as adjusted mean [SE] and are adjusted for the value at 3 months postpartum. All other continuous data are shown as median [interquartile range]. Categorical data are shown as percentages. P values refer to differences between the two groups as determined by Kruskal-Wallis test for continuous variables and either χ2 test or Fisher exact test for categorical variables. AUCgluc, area under the glucose curve; IFG, impaired fasting glucose; IGT, impaired glucose tolerance; NGT, normal glucose tolerance.

Table 1 shows the comparison of characteristics at 3 and 12 months postpartum between women in whom β-cell function (ISSI-2) declined over this time (decliners; n = 92) and those in whom it did not (nondecliners; n = 74). Lipids, adipokines, and C-reactive protein are shown in Supplementary Table 1. Although glucose measures did not differ between the groups at 3 months, each postload glucose value on the OGTT was higher in the decliners than in the nondecliners by 12 months postpartum. Matsuda index, 1/HOMA-IR, fasting C-peptide:insulin, and ISSI-2 were all lower in the decliners at 12 months, as was the prevalence of breastfeeding. Although decliners lost less weight, the two groups did not differ in their baseline-adjusted changes in waist, BMI, and whole-body insulin sensitivity (Matsuda) between 3 and 12 months. In contrast, however, both measures of hepatic insulin sensitivity (1/HOMA-IR and fasting C-peptide:insulin) decreased in decliners but increased in nondecliners.

On Spearman correlation analysis (Supplementary Table 2), the factor most strongly associated with the baseline-adjusted change in ISSI-2 between 3 and 12 months was the change in 1/HOMA-IR (r = 0.45; P < 0.0001). Similarly, the change in fasting C-peptide:insulin was also a significant correlate of the change in ISSI-2 (r = 0.22; P = 0.008), whereas the change in whole-body insulin sensitivity (Matsuda) did not reach significance (r = 0.15; P = 0.08).

On multiple linear regression analyses (Supplementary Table 3) adjusted for age, ethnicity, family history of diabetes, breastfeeding, BMI, and ISSI-2 at 3 months, the only independent determinant of the change in ISSI-2 was the change in hepatic insulin sensitivity, measured by either 1/HOMA-IR (t = 5.5; P < 0.0001) or fasting C-peptide:insulin (t = 3.24; P = 0.0015). The changes in weight or Matsuda index were not significant predictors.

On logistic regression analysis (Supplementary Table 4), an increase in 1/HOMA-IR independently predicted a lower likelihood of declining β-cell function (OR = 0.13 [95% CI 0.06–0.29]; P < 0.0001), after adjustment for age, ethnicity, family history of diabetes, breastfeeding, and BMI at 3 months. Similarly, an increase in fasting C-peptide:insulin also predicted a lower risk of declining β-cell function (OR = 0.95 [0.91–0.98]; P = 0.0018), whereas the changes in weight and Matsuda index were not significant predictors.

Although studies beginning at 15–30 months postpartum have shown that insulin resistance contributes to declining β-cell function after GDM (3,4), the current data specifically implicate hepatic insulin resistance as relevant to the earlier deterioration of β-cell function in the first year postpartum. Interestingly, Buchanan and colleagues (14) have demonstrated that increased endogenous glucose production in pregnancy (indicative of hepatic insulin resistance) independently predicts the subsequent development of diabetes by 11–26 months postpartum. The current findings thus may reflect the early postpartum continuation of this chronic hepatic insulin resistance in those women in whom β-cell function deteriorates, leading to type 2 diabetes. Indeed, this model may be further supported by the demonstration of altered hepatic lipid storage as an early abnormality in women with GDM, even when maintaining normal glucose tolerance (15). It should also be noted that, when compared with nondecliners, the decliners had slightly higher hepatic insulin sensitivity at 3 months postpartum, which then deteriorated over the next 9 months, accompanied by worsening glycemia. This observation suggests that consideration of the pattern of change in hepatic insulin sensitivity over time may be more informative than a single one-time measurement.

Limitations of this study include the observational design that precludes definitive commentary on causality and the use of surrogate indexes. Nevertheless, these data raise the possibility that worsening hepatic insulin resistance may provide an early marker for identifying those women with GDM/GIGT who are most likely to progress to prediabetes/diabetes. Further study is thus needed to determine whether the liver or aspects of its physiology could provide a target for risk stratification and possibly modification in this high-risk patient population.

This study was supported by operating grants from the Canadian Institutes of Health Research (CIHR; MOP-67063 and MOP-84206), the Canadian Diabetes Association (CDA; OG-3-08-2543-RR), and the Heart and Stroke Foundation of Ontario (NA 6747). R.R. holds a CIHR Clinical Research Initiative New Investigator Award and CDA Clinician-Scientist incentive funding. A.J.G.H. holds a Tier-II Canada Research Chair in Diabetes Epidemiology. B.Z. holds the Sam and Judy Pencer Family Chair in Diabetes Research at Mount Sinai Hospital and University of Toronto.

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

R.R. designed the analysis plan, researched data, wrote the manuscript, and was involved in the design and implementation of the overall study. Y.Q. and C.Y. performed the statistical analyses. A.J.G.H., P.W.C., M.S., and B.Z. were involved in the design and implementation of the overall study.

Parts of this study were presented in abstract form at the 71st Scientific Sessions of the American Diabetes Association, San Diego, California, 24–28 June 2011.

The authors thank Mount Sinai Hospital Department of Pathology and Laboratory Medicine and Patient Care Services.

1.
Buchanan
TA
,
Xiang
AH
,
Kjos
SL
,
Watanabe
R
.
What is gestational diabetes?
Diabetes Care
2007
;
30
(
Suppl. 2
):
S105
S111
[PubMed]
2.
Retnakaran
R
,
Qi
Y
,
Sermer
M
,
Connelly
PW
,
Hanley
AJ
,
Zinman
B
.
Glucose intolerance in pregnancy and future risk of pre-diabetes or diabetes
.
Diabetes Care
2008
;
31
:
2026
2031
[PubMed]
3.
Xiang
AH
,
Kawakubo
M
,
Trigo
E
,
Kjos
SL
,
Buchanan
TA
.
Declining β-cell compensation for insulin resistance in Hispanic women with recent gestational diabetes mellitus: association with changes in weight, adiponectin, and C-reactive protein
.
Diabetes Care
2010
;
33
:
396
401
[PubMed]
4.
Xiang
AH
,
Kjos
SL
,
Takayanagi
M
,
Trigo
E
,
Buchanan
TA
.
Detailed physiological characterization of the development of type 2 diabetes in Hispanic women with prior gestational diabetes mellitus
.
Diabetes
2010
;
59
:
2625
2630
[PubMed]
5.
Retnakaran
R
,
Qi
Y
,
Sermer
M
,
Connelly
PW
,
Hanley
AJ
,
Zinman
B
.
β-Cell function declines within the first year postpartum in women with recent glucose intolerance in pregnancy
.
Diabetes Care
2010
;
33
:
1798
1804
[PubMed]
6.
Retnakaran
R
,
Qi
Y
,
Connelly
PW
,
Sermer
M
,
Hanley
AJ
,
Zinman
B
.
Risk of early progression to prediabetes or diabetes in women with recent gestational dysglycaemia but normal glucose tolerance at 3-month postpartum
.
Clin Endocrinol (Oxf)
2010
;
73
:
476
483
[PubMed]
7.
Hoffman
RP
.
Indices of insulin action calculated from fasting glucose and insulin reflect hepatic, not peripheral, insulin sensitivity in African-American and Caucasian adolescents
.
Pediatr Diabetes
2008
;
9
:
57
61
[PubMed]
8.
Tripathy
D
,
Almgren
P
,
Tuomi
T
,
Groop
L
.
Contribution of insulin-stimulated glucose uptake and basal hepatic insulin sensitivity to surrogate measures of insulin sensitivity
.
Diabetes Care
2004
;
27
:
2204
2210
[PubMed]
9.
Hoffman
RP
,
Vicini
P
,
Cobelli
C
.
Pubertal changes in HOMA and QUICKI: relationship to hepatic and peripheral insulin sensitivity
.
Pediatr Diabetes
2004
;
5
:
122
125
[PubMed]
10.
Polonsky
KS
,
Rubenstein
AH
.
C-peptide as a measure of the secretion and hepatic extraction of insulin. Pitfalls and limitations
.
Diabetes
1984
;
33
:
486
494
[PubMed]
11.
Gower
BA
,
Granger
WM
,
Franklin
F
,
Shewchuk
RM
,
Goran
MI
.
Contribution of insulin secretion and clearance to glucose-induced insulin concentration in African-American and Caucasian children
.
J Clin Endocrinol Metab
2002
;
87
:
2218
2224
[PubMed]
12.
Retnakaran
R
,
Hanley
AJ
,
Sermer
M
,
Zinman
B
.
The impact of insulin resistance on proinsulin secretion in pregnancy: hyperproinsulinemia is not a feature of gestational diabetes
.
Diabetes Care
2005
;
28
:
2710
2715
[PubMed]
13.
Meier
JJ
,
Holst
JJ
,
Schmidt
WE
,
Nauck
MA
.
Reduction of hepatic insulin clearance after oral glucose ingestion is not mediated by glucagon-like peptide 1 or gastric inhibitory polypeptide in humans
.
Am J Physiol Endocrinol Metab
2007
;
293
:
E849
E856
[PubMed]
14.
Buchanan
TA
,
Xiang
AH
,
Kjos
SL
,
Trigo
E
,
Lee
WP
,
Peters
RK
.
Antepartum predictors of the development of type 2 diabetes in Latino women 11-26 months after pregnancies complicated by gestational diabetes
.
Diabetes
1999
;
48
:
2430
2436
[PubMed]
15.
Prikoszovich
T
,
Winzer
C
,
Schmid
AI
, et al
.
Body and liver fat mass rather than muscle mitochondrial function determine glucose metabolism in women with a history of gestational diabetes mellitus
.
Diabetes Care
2011
;
34
:
430
436
[PubMed]

Supplementary data