Body surface area: a new anthropometric measurement for diabetic retinopathy in Chinese adults with type 2 diabetes: a hospital registry case-control study

To the Editor: Diabetic retinopathy (DR) is a condition characterized by retinal microvascular damages of diabetes mellitus. As the primary cause of blindness in adults, DR has already become a public health problem worldwide and threatens the outcomes of diabetic patients and even healthy population. The correlation between DR and common anthropometric measurements represented by body mass index (BMI) was still controversial.^[1] Another basic anthropometric index also calculated by height and weight, body surface area (BSA) could more accurately reflect the size of body, and distinguish adipose or muscle tissues.^[2] In addition, previous research considered that adipokines secreted from adipose tissue might play an important role in the mechanism of obesity and type 2 diabetes. The level of adiponectin decreased in obese people and lower adiponectin might cause insulin resistance and diabetes.^[3] We firstly investigated the relationship between BSA and DR in Chinese type 2 diabetic patients, and compared it with BMI. The study was approved by the Institutional Review Board of the First Affiliated Hospital of China Medical University (AF-SOP-07-1.1-01/2019-13). Written informed consent was obtained for the entire study population included in this study.

This hospital registry case-control study involved 2454 Chinese subjects (age ≥18 years) who were type 2 diabetic inpatients. The selection procedure of cases and controls was in accordance with Supplementary Figure 1. Type 2 diabetes was defined following American Diabetes Association standards. DR was evaluated according to Early Treatment Diabetic Retinopathy Study and International Classification Diabetic Retinopathy Scales,^[4] which was categorized in four severities as non-DR, mild non-proliferative DR (NPDR), moderate NPDR, and vision-threaten DR (VTDR). BMI was calculated as dividing weight (kg) by squared height (m²). BSA was calculated by 0.007184 × weight (kg)^0.425 × height (cm)^0.725.^[5] All analysis was performed by SPSS version 22 (IBM Corp, Armonk, NY, USA). P < 0.05 was considered statistically significant.

The Table 1 summarizes the basic characteristics of each participant through collecting the results of detailed assessments at the time of his/her initial admission. One-way analysis of variance test and Chi-squared test were applied to explore the correlations between variables and DR severity. Among all variables, duration of diabetes, heart rate, systolic blood pressure, pulse pressure, weight, BMI, and BSA levels were associated with DR severity significantly (all P < 0.05). The correlation P value between any-severity DR and BSA was 0.025 (BSA of DR 1.77 ± 0.18 vs. BSA of non-DR 1.79 ± 0.19), but for any-severity DR and BMI, the correlation had no significance (P = 0.616). Regardless of gender (P = 0.008) or only including males (P < 0.001), BSA level tended to decrease with the advanced DR severity significantly, but not existed in females. Due to the body fat distribution was discrepant in different genders, the results of males were more convincing without the interference of females.

Table 1.

Characteristics of study participants by the presence and severity of DR.

	Non-DR	Mild NPDR	Moderate NPDR	VTDR
Variables	(n = 1636)	(n = 609)	(n = 108)	(n = 101)	P values
Male, n (%)	910 (55.6)	345 (56.7)	54 (50.0)	56 (55.4)	0.649
Age (years)	57.41 ± 12.33	57.20 ± 12.16	59.56 ± 12.99	56.42 ± 12.52	0.247
Duration of diabetes, n (%)
≤5.0 years	749 (45.8)	225 (36.9)	26 (24.1)	41 (40.6)	<0.001
5.1–10.0 years	403 (24.6)	155 (25.5)	29 (26.9)	25 (24.8)
10.1–20.0 years	404 (24.7)	191 (31.4)	41 (38.0)	30 (29.7)
>20.0 years	80 (4.9)	38 (6.2)	12 (11.1)	5 (5.0)
Anti-diabetic agents∗ (yes), n (%)	1529 (93.5)	582 (95.6)	105 (97.2)	95 (94.1)	0.138
HR (beats/min)	80.00 ± 10.42	81.82 ± 11.70	80.78 ± 11.84	83.84 ± 10.15	<0.001
SBP (mmHg)	133.26 ± 17.83	136.67 ± 19.33	136.33 ± 20.78	132.68 ± 19.03	0.001
DBP (mmHg)	82.19 ± 10.86	82.97 ± 11.59	81.44 ± 10.10	81.83 ± 12.57	0.369
PP (mmHg)	51.07 ± 14.94	53.70 ± 15.80	54.89 ± 18.29	50.85 ± 14.78	0.001
Height (cm)	167.79 ± 8.10	167.24 ± 8.10	166.59 ± 9.01	166.46 ± 7.77	0.129
Weight (kg)	69.73 ± 12.52	69.30 ± 12.65	68.17 ± 11.92	65.25 ± 11.27	0.004
BMI (kg/m2)	24.65 ± 3.30	24.67 ± 3.47	24.49 ± 3.30	23.53 ± 3.63	0.012
Male	25.27 ± 3.30	25.05 ± 3.47	24.71 ± 3.39	23.62 ± 3.75	0.003
Female	23.87 ± 3.14	24.18 ± 3.42	24.26 ± 3.22	23.42 ± 3.52	0.326
BSA (m2)	1.79 ± 0.19	1.78 ± 0.19	1.76 ± 0.18	1.72 ± 0.16	0.008
Male	1.90 ± 0.15	1.88 ± 0.16	1.87 ± 0.16	1.81 ± 0.13	<0.001
Female	1.65 ± 0.13	1.65 ± 0.13	1.65 ± 0.13	1.62 ± 0.14	0.635
Smoker∗ (yes), n (%)	435 (26.6)	139 (22.8)	25 (23.1)	31 (30.7)	0.171
Drinker∗ (yes), n (%)	318 (19.4)	119 (19.5)	25 (23.1)	20 (19.8)	0.828
HbA1c, % (mmol/mol)	8.44 ± 2.27 (68.7 ± 24.9)	8.59 ± 2.75 (70.4 ± 30.0)	8.40 ± 2.40 (68.3 ± 26.1)	8.75 ± 2.46 (72.2 ± 26.8)	0.344
FPG (mmol/L)	9.25 ± 3.83	9.23 ± 3.89	9.03 ± 3.16	9.78 ± 4.11	0.520

^∗Definition: Either oral hypoglycemic drugs or insulin was thought to be in use of anti-diabetic agents. Smoker or drinker was considered as self-described past or current history of smoke or drink no matter the amount.

DR: Diabetic retinopathy; NPDR: Non-proliferative diabetic retinopathy; VTDR: Vision-threatening diabetic retinopathy; HR: Heart rate; SBP: Systolic blood pressure; DBP: Diastolic blood pressure; PP: Pulse pressure; BMI: Body mass index; BSA: Body surface area; HbA1c: Glycated hemoglobin; FPG: Fasting plasma glucose.

Next, Supplementary Table 1 performed receiver operating characteristic curves to estimate the prediction value of BMI and BSA, the results depended on the area under curve (AUC). BSA levels were significant in predicting VTDR in all population (AUC = 0.591, P = 0.002) or only males (AUC = 0.674, P < 0.001). BMI also had the similar significances in predicting VTDR. However, in the prediction of any-severity DR, only BSA had notable predictive effects (all population: AUC = 0.527, P = 0.027; only male: AUC = 0.552, P = 0.002). Comprehensively compare the predictive ability of BSA and BMI to DR, BSA was more pronounced in all DR levels. The findings suggested that BSA was more closely related to DR and its severity, either in the whole population or only in males.

In addition, BSA was divided into four categories by its quartiles and represented by the letter Q. Variables were compared according to BSA quartiles in Supplementary Table 2. The trend of BSA quartiles and DR severity in different genders was shown in Supplementary Figure 2. Further, this study designed three multi-variable-adjusted models to investigate the associations between DR severity and BSA level using logistic regression, setting non-DR as reference. In Supplementary Table 3, after adjusting for age and gender in model 1, BSA was associated with any-severity DR (odds ratio 0.40, 95% confidence interval 0.22–0.74; P = 0.003) and VTDR (0.05, 0.01–0.21; P < 0.001) significantly, but not with mild NPDR or moderate NPDR. Further adjusted in model 3, the associations between BSA and any-severity DR (0.40, 0.22–0.74; P = 0.003) as well as VTDR (0.06, 0.01–0.23; P < 0.001) were still significant. Included male participants only in Supplementary Table 4, the results of logistic regression were more meaningful in any-severity DR and VTDR group (all P ≤ 0.001). Moreover, the similar association in males was also significant in the mild NPDR group.

In Supplementary Table 5, BSA quartile was regarded as categorized variables and set Q1 as reference. When applied logistic regressions in model 3, there was a remarkable dose-response trend between BSA quartiles and any-severity DR (P = 0.006), and VTDR (P = 0.001), suggesting a negative correlation between BSA quartiles and the risk of DR and VTDR. Supplementary Table 6 included males only, similar dose-response relationships were also statistically significant in mild NPDR (P = 0.025), in other words, odds ratio decreased with the growth of BSA quartiles.

Up to now, few researches have proved the relationship of BSA level with diabetic microangiopathy. Our study showed that BSA correlated with DR severity reversely better than BMI, proving to be an independent DR predictor even among Chinese adults with type 2 diabetes. However, this association should be with cautious in female. Despite these, there was no conclusion that it was necessary to initiatively improve BMI or BSA levels to prevent DR. More prospective studies, not just for hospitalized patients, were needed to support the benefits of higher BSA levels for DR.

Acknowledgements

The authors thank all study participants who contributed data to our research. They also thank other investigators in the Department of Endocrinology and Metabolism, First Hospital of China Medical University.

Funding

This work was supported by grants from the National Natural Science Foundation of China (No. 81300783), China Postdoctoral Science Foundation (No. 2019TQ0358; No. 2019M661162), Shenyang Young and Middle-aged Science and Technology Innovation Talent Support Program (No. RC190146), and Liaoning Revitalization Talents Program (No. XLYC1807082).

Conflicts of interest

None.