3.1 Demographic and biochemical characteristics of cases and controls
From June 2015 to June 2019, 263 women with PCOS and 326 controls were included in this study. Demographic characteristics and biochemical parameters were summarized in Table 1. Significant differences were found between PCOS cases and controls in BMI, FSH, LH, LH/FSH, and T (P < 0.001 for all comparisons), except for E2 (P = 0.07).
Table 1. Clinical characteristics of the women with PCOS and the control subjects.
Variable
|
Control (n = 326)
|
PCOS (n = 263)
|
P-valuea
|
Age (y)
|
29.00 (28.00, 33.00)
|
29.00 (27.50, 31.00)
|
0.133
|
BMI (kg/m2)
|
20.80 (19.49, 22.86)
|
23.80 (21.69, 25.96)
|
<0.001
|
FSH (IU/L)
|
7.70 (6.40, 8.96)
|
6.80 (5.70, 7.90)
|
<0.001
|
LH (IU/L)
|
4.20 (3.10, 5. 30)
|
7.90 (4.90, 12.40)
|
<0.001
|
LH/FSH
|
0.53 (0.40, 0.71)
|
1.16 (0.72, 1.75)
|
<0.001
|
E2 (pmol/L)
|
164.00 (119.20, 206.80)
|
173.00 (125.50, 230.50)
|
0.07
|
T (nmol/L)
|
1.41 ± 0.46
|
2.24 ± 0.76
|
<0.001
|
Data are given as median (IQR) or mean ± SD.
IQR, interquartile range. SD, standard deviation.
a p values obtained by comparison of variables between PCOS and controls using the Mann–Whitney U-test.
3.2 Increased relative mtDNA CN and mtDNA4977 DR in PCOS
The mtDNA CN and mtDNA4977 DR between the two groups were compared using the Mann–Whitney U‐test and the Student’s t-test, respectively. As shown in Figure 1, the mean log-transformed mtDNA CN in women with PCOS was statistically higher than that in the controls (2.00 versus 1.95, P = 0.002). Additionally, the mtDNA4977 DR was significantly higher in the PCOS patients compared with controls (1.52 versus 1.48, P < 0.001).
Further, all subjects were divided into four groups according to the quartiles of LH/FSH. A significant difference was observed in mtDNA CN among the four groups, as well as in mtDNA4977 DR (Supplementary Table S2). After Bonferroni’s correction, a significant difference was confirmed between the lowest quartile and highest quartile (P = 0.011) for mtDNA4977 DR. However, no differences were found between the four groups in terms of mtDNA CN.
3.3 Associations of mtDNA CN and mtDNA4977 DR with PCOS
The associations of mtDNA CN and mtDNA4977 DR with PCOS were assessed using logistic regression models (Table 2). The mtDNA CN was significantly associated with PCOS. However, after adjustment for BMI, LH/FSH, and T, mtDNA CN as a linear continuous variable was not associated with PCOS (OR = 1.004, 95% CI: 0.999, 1.008). It was consistent with the quartile analyses results, which showed no statistically significant associations between the first quartile and higher quartiles (Figure 2, Supplementary Table S3).
Table 2. Odds ratios and 95% CIs for PCOS with mtDNA biomarkers and metabolic parameters.
|
Odds ratio (95% CI)
|
P-value
|
Adjusted† OR
(95% CI)
|
P-value
|
mtDNA CN
|
1.005 (1.002, 1.008)
|
0.003
|
1.004 (0.999, 1.008)
|
0.091
|
mtDNA4977 DR
|
1.047 (1.027, 1.069)
|
< 0.001
|
1.053 (1.024, 1.083)
|
< 0.001
|
mtDNA, mitochondrial DNA; mtDNA4977, mitochondrial DNA 4977-bp; CN, copy number; DR, deletion rate; OR, odds ratio; CI, confidence interval.
†Adjusted for body mass index, luteinizing hormone/follicle-stimulating hormone, and testosterone.
Similarly, mtDNA4977 DR was significantly associated with PCOS (P < 0.001). However, unlike mtDNA CN, the linear dose-response trends were still apparent in the adjusted model. As can be observed in Table 2, the adjusted OR (95% CI) for PCOS by mtDNA4977 DR was 1.053 (1.024, 1.083). This relationship was also supported by the quartile analysis results (Figure 2, Supplementary Table S3). Compared to the first quartile, the higher quartiles of mtDNA4977 DR were associated higher risk of PCOS, especially in the highest quartile (OR = 3.867, 95% CI: 1.847, 8.096). These results imply that the association between mtDNA4977 DR and PCOS is independent of metabolic parameters.