Impaired Respiratory Function in Women With PCOS Compared With Matched Controls From a Population-Based Study

Abstract

Context

Increased prevalence of asthma has been reported from epidemiological studies in women with polycystic ovary syndrome (PCOS).

Objective

To investigate respiratory function in women with PCOS compared with controls in a clinical setting.

Design

An 8-year clinical follow-up study including self-reported asthma diagnoses and spirometry of women with PCOS randomized to metformin or placebo during pregnancy in the original studies (the Pilot and the PregMet-study), compared with matched controls from a population-based cohort study (The HUNT Study).

Setting

Secondary and tertiary care centers.

Participants

A total of 145 women with PCOS (54% of original cohort) were matched 1:3 to controls, on gender, age, and smoking-status.

Main outcomes and measures

Self-reported doctor-diagnosed asthma (DDA), percentage of predicted forced expiratory volume in the first second of expiration (FEV1 % predicted), percentage of predicted forced vital capacity (FVC % predicted).

Results

Women with PCOS reported more DDA compared with controls (19% vs 9%; P < 0.01). Spirometry indicated a combined obstructive (FEV1 % predicted, 93.7 vs 102.0; P < 0.01) and restrictive (FVC % predicted, 94.5 vs 103.7; P < 0.01) respiratory impairment in PCOS compared with controls. Metformin in pregnancy did not affect respiratory function at follow-up.

Conclusion

Women with PCOS reported higher prevalence of DDA compared with controls matched for age and smoking status. In addition, respiratory function was decreased, with both obstructive and restrictive components. Further insight to the underlying pathogenesis of these observations is needed.

Clinical trial registration

ClinicalTrials.gov: The PregMet study: NCT00159536. The Pilot study: NCT03259919.

Follow-up, asthma, spirometry, metformin, PCOS, HUNT

Polycystic ovary syndrome (PCOS) is the most prevalent endocrine condition in women of fertile age (1). Although the etiology is multifactorial and only partly known, women with a genetic predisposition exposed to an obesogenic environment are at an increased risk of developing the syndrome (2). Reproductive and metabolic disturbances are the most frequent implications of PCOS, whereas disorders related to immune system impairment, such as respiratory, thyroid, and rheumatic disorders, have gained less attention (3). Register-based data from Denmark indicated that women with PCOS had higher prevalence of an asthma diagnosis at hospital discharge and that prescriptions of asthma medication were higher in PCOS patients than in age-matched controls: 3.0% vs 2.2% and 19.2 vs 14.1% respectively (4).

Asthma is a heterogeneous disease, and descriptions of asthma in both children and adults commonly include airway hyperresponsiveness and airway inflammation as components of the disease (5). Central to all definitions is the presence of symptoms (> 1 of wheezing, breathlessness, chest tightness, coughing) and of variable airflow obstruction (6). In childhood, asthma is more prevalent and severe in boys, but around puberty, asthma becomes more prevalent in girls and prevalence increases with increasing Tanner stages in girls (7). This phenomenon of gender reversal showed consistency across populations in data from the European Community Respiratory Health Survey II (8). Gender differences in maternal cortisol levels during the third trimester may protect female offspring from early asthma, as maternal cortisol levels are significantly higher in mothers carrying a female offspring (9).

PCOS is closely associated with overweight and obesity. Obesity, especially central adiposity, was associated with asthma in women in an epidemiological study (10). This raises the question whether the observed increase of asthma in PCOS primarily is linked to obesity, or if the key drivers of PCOS—insulin resistance and hyperandrogenism—further increase the risk of asthma in PCOS (11, 12). In an Australian cohort, the increased prevalence of self-reported asthma in PCOS remained significant after adjustment for age, body mass index (BMI), and smoking (13); this may indicate an independent effect of PCOS on asthma. Additionally, an association between impaired glucose regulation and decreased lung function has been described in nondiabetic populations (14, 15). In a large cross-sectional study, oligomenorrhea, 1 of 3 criteria for PCOS diagnosis, was associated with increased prevalence of asthma symptoms and decreased forced vital capacity (FVC) at all values of BMI above 25 kg/m² (16). PCOS, inferior lung function, and increased prevalence of asthma have all been associated with an early age at puberty (17–20).

Inflammatory markers are elevated in both lean and obese women with PCOS and inflammatory mechanisms are well documented in asthma (21). In a murine model of chronic asthma, metformin suppressed eosinophilic inflammation and reduced airway remodeling (22). Although not used primarily for asthma, treatment with metformin in type 2 diabetes was associated with a lower incidence of asthma in a retrospective cohort study (23).

The aim of the present study was to investigate the prevalence of asthma and explore respiratory function in women with PCOS versus matched controls, based on an internationally validated questionnaire and spirometry. As PCOS participants were randomized to metformin or placebo from the first trimester and onwards in pregnancy, we also aimed to explore whether metformin during pregnancy affected long-term respiratory health at 8-year follow-up.

Materials and Methods

Participants and study design

PCOS participants.

Participants in the current study were originally participants from 2 previous studies: a pilot study (24) and the PregMet study (25). Inclusion criteria for these studies were essentially identical: 1) PCOS diagnosed according to the Rotterdam criteria (26), 2) age 18–45 years, 3) gestational age of fetus between 5 and 12 weeks and 4) a singleton viable fetus. The pilot study included 40 women and investigated the effect of metformin on androgen levels in pregnant women with PCOS (24). Participants were randomized to metformin 1700 mg daily or placebo from the first trimester until delivery.

The PregMet study included 274 pregnancies and explored the efficacy of metformin to reduce pregnancy complications in women with PCOS (25). In the PregMet study, participants were randomized to metformin 2000 mg daily or placebo until delivery.

Participants gave their consent to be contacted after delivery for possible follow-up studies. Of the 314 pregnancies included in the pilot and the PregMet studies, 1 participant was excluded due to misdiagnosis, 14 dropped out, 4 miscarried, and 3 children died in the follow-up period and their mothers were not approached at follow-up. Seventeen women participated twice in the PregMet study and 8 women participated once in the pilot and once in the PregMet study (Fig. 1 flowchart).

Figure 1.

Flowchart of the original and follow-up studies.

Follow-up of PCOS participants.

Follow-up of invited participants (N = 268) was carried out between April 2014 and July 2016. In all, 145 women (54%) agreed to participate in the follow-up; 131 women met for physical examination and interview, whereas 14 were interviewed by phone and gave self-reported data. Standardized interviewer-administered questionnaires were used to obtain self-reported data on former medical history, smoking status and smoking—in pack-years—as a numerical value of lifetime tobacco exposure, contraceptives, physical activity, and education. A separate form covered previous respiratory diseases, symptoms of respiratory disease such as wheezing in the last 12 months, breathlessness at night, daily cough during periods of the year, and atopy (hay fever/nasal allergies). In cases of reported asthma, participants were asked to specify whether the asthma diagnosis was received after evaluation and examination by a physician, later labeled “doctor-diagnosed asthma” (DDA). The term “self-reported asthma” included all cases of reported asthma, including cases where a physician-confirmed diagnosis was not clearly specified by the participant. These questions were based on The European Community Respiratory Health Survey (ECRHS) II (27). Participants were asked to tick off boxes of drawn body silhouettes on a scale 1 (very lean) to 9 (extremely obese) which best described their figure at the age of 8 years, at menarche, and currently. In line with a previous validation study, cutoff at figural scale ≥ 4 was optimal in detecting overweight and/or waist circumference > 88 cm (28). Height and waist circumferences were measured manually and rounded off to the closest 0.5 cm. Weight was measured using bioelectrical impedance (Inbody 720, BIOSPACE, Seoul, Korea) for women who met for examination, or manually for women who gave self-reported data. Baseline data from the first participation were recorded in the cases where a woman participated twice in the original randomized controlled trials and randomization was equal in both pregnancies; if randomized once to metformin and once to placebo, data from the metformin-exposed pregnancy were used. None of the women were postmenopausal when included in the follow-up. Contraceptives were not discontinued before the follow-up examination.

Control participants from the HUNT study.

The Nord-Trøndelag Health Study (The HUNT Study) is an ongoing population-based cohort study inviting all residents, 20 years or older (N = 93 869), in Nord-Trøndelag County of Norway to attend health surveys with clinical interviews and examinations. Three surveys have been completed, HUNT1 (1984–86), HUNT2 (1995–97), and HUNT3 (2006–08) (29). Controls from HUNT3 were surveyed at approximately the same timeframe as the follow-up of women with PCOS was conducted. In HUNT3, 54.1% of the invited population participated. The HUNT3 cohort profile has been published previously (29). The HUNT3 Survey included questionnaires, interview and measurements at the screening stations. Questions from ECRHS II were embedded in the HUNT3 Lung Survey. Height and weight were measured with the participants wearing light clothes without shoes and given in centimeters/kilograms with one decimal. Waist circumference was measured at the height of the umbilicus to the nearest 1.0 cm, with the participant standing straight with her arms hanging relaxed. From HUNT3, data for 3 control women were extracted based on the best possible match on age and smoking status (current vs noncurrent smoker) for each woman with PCOS. Of 2286 possible controls, women stating current pregnancy (N = 18), oligomenorrhea (N = 9) and menopause (N = 1017) were excluded before matching.

Laboratory analyses in PCOS participants

Venous blood samples were collected and processed after an overnight fast at each study site. Without delay, serum was analyzed for glucose on an Advia Chemistry XPT at St Olavs Hospital, Trondheim, Norway. For other analyses, serum was stored at −80°C, and thawed once before analysis. Serum was analyzed for insulin by electrochemiluminescence immunoassay at Aker, Oslo University Hospital, Norway. Homeostatic model assessment for insulin resistance (HOMA-IR) index was computed as (fasting glucose mmol/L x fasting insulin concentration mU/L)/22.5. Adiponectin was analyzed by radioimmunoassay at Aker, Oslo University Hospital, Norway. Testosterone was analyzed by liquid chromatography-tandem mass spectrometry (LC-MS/MS) at the Department of Clinical Chemistry, Vejle County Hospital, Denmark. In the analysis, plasma samples were extracted by supported liquid extraction and the eluate evaporated and reconstituted before analysis on LC-MS/MS. The analysis was calibrated by in-house prepared calibrators and the relative standard deviation was below 10%. Quality was assured by monthly participation with satisfactory results in the external quality control program for steroid hormones from NEQAS, UK. Sex hormone-binding globulin (SHBG) was analyzed by noncompetitive immunoluminometric assay (ILMA) at Aker. Free Testosterone Index was calculated as (total testosterone/SHBG) x 100. Calculated free testosterone or calculated bioavailable testosterone was calculated as described by Vermeulen et al (30).

Spirometry

Spirometry in the follow-up study (EasyOne spirometer, ndd Medical Technologies, Chelmsford, Massachusetts) and in the HUNT3 Survey (MasterScope Jaeger version 5.1; JAEGER, Wuerzburg, Germany) was conducted in accordance with the American Thoracic Society/European Respiratory Society recommendations (31). When compared with laboratory spirometry, EasyOne handheld spirometer has been regarded as valid and reproducible in both clinical and research settings (32). Predicted values were calculated using the Global Lung Function Initiative 2012 (GLI-2012) software (33). Forced vital capacity (FVC), is the maximal total volume of air exhaled by maximally forced effort after a full inspiration. Absolute forced expiratory volume in one second (FEV1) is a measure of volume exhaled in the first second of the FVC maneuver. As lung function is dependent on age, height, and ethnicity, spirometry results are presented as percentage of predicted values, FVC % predicted and FEV1 % predicted, respectively. Reduced FVC % predicted indicates restrictive pulmonary conditions, and reduced FEV1 reflects obstruction of the airways, used in monitoring of asthma. Measurements of airway reversibility by bronchodilator, often used in diagnosing asthma, was not evaluated as required in the current setting.

Statistical analyses

Each case was matched to 3 controls with the same current smoking status (current smoking yes/no) and age. Medians with 95% CI, or counts and proportions were computed for outcome variables for cases and controls as appropriate.

Cases and controls were compared using either a permutation test, a mixed logistic test or the Mann-Whitney U-test as indicated in the tables. For the permutation test, each case’s rank within its matched group were summed across cases. The permutation test was used for continuous and ordinal variables when there was complete data in each matched group for most of the groups. The mixed logistic test was used for dichotomous variables. In the mixed model, case vs control was the fixed effect and matched group ID was the random effect. Both the permutation test and the mixed logistic test accounted for the dependence between observations due to the matching. The Mann-Whitney U-test was used for continuous and ordinal variables when there was substantial missing data in the matched groups.

Predictors of DDA for the cases was examined using logistic regression.

The statistical analyses were done using R version 2.13.1 (R Development Core Team, Vienna, Austria) and IBM SPSS Statistics version 25.0 (IBM, SPSS Inc USA, Chigaco IL). The level of significance was taken as 0.05.

Ethical approval

Written informed consent was obtained from each participant before inclusion in the present follow-up and the declaration of Helsinki was followed throughout the study. The Regional Committee for Health Research Ethics of Central Norway approved the present study 04.04.2014, reference number: 2014/96.

Results

In all, 145 of 268 invited women with PCOS (54%) agreed to participate in this follow-up, all diagnosed in accordance with the Rotterdam criteria, of whom 107/145 (74%) had a hyperandrogenic phenotype at inclusion into the original PregMet study (data not shown).

Women with PCOS had higher BMI (median 28.5 vs 25.7 kg/m²; P ≤ .01) and higher level of physical activity and education compared with controls (Table 1). The prevalence of reported menarche < 13 years was higher in women with PCOS.

Table 1.

Anthropometric, Clinical, and Socioeconomic Characteristics of Women With PCOS and Matched Control Women

	N	PCOS*	N	Control	P Value
Age (years)	145	38 (37–39)	435	38 (37–40)
Smoking current (%)	145	13 (9)	435	39 (9)
Smoking former (%)	145	46 (35)	435	140 (35)	1.0b
Smoking (pack-years)	59	5.0 (2.5–7.0)	179	4.3 (3.8–5.6)	1.0c
Height (cm)	145	167 (166–168)	432	167 (166–168)	0.71a
BMI (kg/m2)	145	28.5 (26.7–30.0)	432	25.7 (25.2–26.3)	<0.01a
BMI (kg/m2) in categories	145		432		<0.01a
BMI ≤ 24.9 (%)		40 (28)		195 (45)
BMI = 25.0–29.9, overweight (%)		44 (30)		155 (36)
BMI ≥ 30.0, obesity (%)		61 (42)		82 (19)
Waist (cm)	144	90 (86–94)	432	88 (86–90)	0.14a
Family history of asthma (%)	145	53 (37)	435	151 (35)	0.69b
Childhood secondhand smoke (%)	145	104 (72)	435	307 (71)	0.79b
OCP use (%)	145	16 (11)	435	46 (11)	0.88b
Menarche <13 years	143	59 (41)	419	131 (31)	0.03b
Physical activity level	144		435		<0.01a
Inactive (%)		14 (10)		64 (15)
Low (%)		26 (18)		97 (22)
Moderate (%)		90 (63)		262 (60)
High (%)		14 (10)		12 (3)
Education completed	143		271		<0.01a
Primary and lower secondary school (%)		8 (6)		103 (38)
Upper secondary school (%)		39 (27)		83 (31)
Higher education /university ≤ 4 years (%)		43 (30)		58 (21)
Higher education/university > 4 years (%)		53 (37)		27 (10)
Currently employed (%)	144	133 (92)	433	402 (93)	0.85b

Notes: Data presented as median (95% CI) or numbers (%) as appropriate. Abbreviations: BMI, body mass index; OCP, oral contraceptive pill; PCOS, polycystic ovary syndrome. *No statistically significant differences (P value < .01) between the metformin and placebo groups at follow-up. ^aPermutation-test. ^bMixed-logistic test ^cMann-Whitney U-test.

Women with PCOS reported more DDA (19% vs 9%; P < 0.01) and self-reported asthma (19% vs 11%; P = 0.02) compared with controls (Table 2). In the PCOS group, women reporting DDA and women with self-reported asthma were identical (asthmatics). Self-reported asthma-related symptoms and atopy were not different in women with PCOS compared with controls. Spirometry indicated a combined obstructive (FEV1 % predicted 93.7 vs 102.0; P < 0.01) and restrictive (FVC % predicted 94.5 vs 103.7; P < 0.01) respiratory impairment in PCOS compared with controls. In both self-reported cases of asthma and nonasthma, the percentage predicted values of FEV1 and FVC were lower in women with PCOS compared with controls (Fig. 2). In sensitivity analyses restricted to nonobese participants, absolute values of FEV1 and FVC were lower in PCOS, and there was a trend towards lower percentage predicted FEV1 and FVC in the PCOS group (Table 2). FEV1 % predicted was not different in asthmatic versus nonasthmatic women with PCOS (median 88.8% vs 95.3%; P = 0.10) (data not shown). Logistic regression analyses did not identify risk-factors for asthma in the PCOS group (Table 3). When comparing treatment of metformin to placebo in women with PCOS, FVC measurements were higher in the metformin-treated group at follow-up (Table 4).

Table 2.

Respiratory Health in Women With PCOS and Matched Control Women

	N	PCOS	N	Control	P Value
Doctor-diagnosed asthma (%)	145	28 (19)	435	38 (9)	<0.01b
Asthma self-reported (%)	145	28 (19)	435	50 (11)	0.02b
Asthma medication, current use (%)	145	11 (8)	435	24 (6)	0.40b
Wheezing last 12 months (%)	145	26 (18)	435	54 (12)	0.09b
Breathlessness at night (%)	145	7 (5)	435	9 (2)	0.17b
Daily cough in periods of the year (%)	145	18 (12)	347	52 (15)	0.46b
Hay fever or nasal allergies (%)	144	37 (26)	430	127 (29)	0.39b
Pulmonary function
FEV1	123	3.01 (2.89–3.08)	337	3.34 (3.22–3.39)	<0.01a
FEV1 % predicted	123	93.7 (90.2–97.2)	337	102.0 (98.3–104.4)	<0.01a
FVC	123	3.68 (3.60–3.78)	334	4.11 (4.01–4.20)	<0.01a
FVC % predicted	123	94.5 (91.3–96.3)	334	103.7 (102.6–107.3)	<0.01a
Analyses in nonobese (BMI <30) women with PCOS and matched control women
BMI (kg/m2)	86	25.0 (24.2–26.1)	351	24.5 (24.0–24.9	0.35c
Doctor-diagnosed asthma (%)	86	20 (23)	351	32 (9)	<0.01d
Asthma self-reported (%)	86	20 (23)	351	40 (11)	<0.01d
Hay fever or nasal allergies (%)	86	24 (28)	351	110 (31)	0.60d
Pulmonary function
FEV 1	71	3.01 (2.89–3.14)	274	3.27 (3.20–3.38)e	<0.01c
FEV1 % predicted	71	93.8 (89.6–97.4)	274	101.8 (100.4–103.3)e	0.10c
FVC	71	3.76 (3.66–3.86)	271	4.16 (4.06–4.21)e	<0.01c
FVC % predicted	71	95.3 (90.8–97.1)	271	105.5 (103.5–107.3)e	0.11c

Notes: Data presented as median (95% CI) or numbers (%) as appropriate. Abbreviations: BMI, body mass index; FEV1, forced expiratory volume in 1 second; FVC, forced vital capacity; PCOS, polycystic ovary syndrome. ^aPermutation-test. ^bMixed-logistic test. ^cMann-Whitney U-test. ^dFisher’s exact test. ^eMatching ignored for computation for CI.

Figure 2.

FEV1 % predicted: PCOS-no asthma (median, 95.2; 95% CI, 91.9–97.4) vs CONTROL-no asthma (median, 102.0; 95% CI, 100.5–103.8); P < 0.01. PCOS-asthma (median, 88.8; 95% CI, 77.1–98.8) vs CONTROL-asthma (median, 97.9; 95% CI, 95.3–102.9); P < 0.01. FVC % predicted: PCOS-no asthma (median, 94.6, 95% CI, 92.3–96.7) vs CONTROL-no asthma (median, 105.2; 95% CI, 103.5–107.4); P < 0.01. PCOS-asthma (median, 91.2; 95% CI, 80.6–100.9) vs CONTROL-asthma (median, 102.2; 95% CI, 100.5–106.2); P < 0.01.

Table 3.

Associations Between Doctor-Diagnosed Asthma and Risk Factors of Asthma in Women With PCOS (N = 28)

	OR	95% CI
Overweight by self-estimated body image at age 8	0.8	0.3–1.8
Overweight by self-estimated body image at menarche	0.7	0.3–1.6
BMI kg/m2, current	1.0	0.9–1.0
Adiponectin	0.9	0.8–1.2
Oligomenore	1.3	0.5–3.1
Hyperandrogenism
T-Testosterone	0.9	0.4–2.0
Free Testosterone Index	1.1	0.8–1.4
Calculated Bioavailable Testosterone (Vermeulen)	1.6	0.2–15.2
FG	1.0	0.9–1.1
Insulin resistance
F-Insulin	1.0	0.9–1.0
F-Glucose	0.7	0.3–1.6
HOMA-index	0.9	0.7–1.2

Notes: Results from logistic regression with nonasthma as reference category. Abbreviations: BMI, body mass index; FG, Ferriman-Gallwey score; HOMA, homeostatic model assessment; OCP, oral contraceptive pill; OR, odds ratio.

Table 4.

Respiratory Health in Women With PCOS According to Study Randomization

	N	Metformin	N	Placebo	P Value
Asthma self-reported (%)	74	11 (15)	71	17 (24)	0.21
Asthma medication, current use (%)	74	6 (8)	71	5 (7)	1
Wheezing last 12 months (%)	74	13 (17)	71	13 (18)	1
Breathlessness at night (%)	74	6 (8)	71	1 (1)	0.12
Daily cough in periods of the year (%)	74	10 (14)	71	8 (11)	0.80
Hay fever or nasal allergies (%)	74	17 (23)	70	20 (29)	0.45
Pulmonary function
FEV1	64	3.07 (2.96–3.14)	59	2.9 (2.77–3.10)	0.14
FEV1 % predicted	64	95.7 (90.0–98.2)	59	92.1 (87.6–97.4)	0.10
FVC	64	3.74 (3.62–3.91)	59	3.61 (3.49–3.78)	0.04
FVC % predicted	64	95.4 (91.3-99-3)	59	93.4 (88.0–95.9)	0.03

Notes: Data presented as median (95% CI) or numbers (%) as appropriate. Abbreviations: FEV1, forced expiratory volume in 1 second; FVC, forced vital capacity

In the peripubertal period (11–15 years), the odds ratio (OR) of a self-reported asthma diagnosis was 8.4 in the PCOS-group compared with controls (P = 0.006) (Fig. 3). In participants with self-reported asthma, breathlessness at night (OR, 6.3) and wheezing (OR, 7.4) were the most frequently reported asthma-symptoms in the PCOS group and the control group, respectively; however, there was no difference in the OR of symptoms between PCOS and controls (Table 5).

Figure 3.

OR = 8.4 (P = 0.006) for asthma in the peripubertal period (11–15 years) in PCOS (black line) versus controls (grey line).

Table 5.

Associations Between Self-Reported Asthma and Different Asthma-Symptoms and Obesity in the PCOS and Control Group

	PCOS (N = 145)		Control (N = 435)		P Value
	OR	95%CI	OR	95%CI
Wheezing last 12 months	5.5	2.2–14.0	7.4	3.6–15.4	0.86
Breathlessness at night	6.3	1.3–30.1	4.0	1.0–16.8	0.67
Daily cough during periods of the yeara	2.4	0.8–7.0	3.3	1.6–6.9	0.63
Obesity, BMI (kg/m2) ≥ 30.0	0.5	(0.2–1.2)	1.1	0.5–2.3	0.18

Notes: Results from mixed logistic regression analysis with self-reported nonasthma as reference category, P values for comparison of odds ratio between POCS and controls. Abbreviations: BMI, body mass index; OR, odds ratio. ^aThe values are based on 145 women with PCOS and 347 control women

Discussion

Women with PCOS had more self-reported asthma than controls. FEV1 % predicted as well as FVC % predicted were lower in the PCOS group compared with controls, suggesting a combined obstructive and restrictive respiratory impairment in women with PCOS. The statistically significant higher FVC-measurements in the metformin-treated group compared with placebo-treated women with PCOS (P < 0.05) was an isolated finding of uncertain clinical importance. We are aware of only 1 other study reporting spirometry measurements in PCOS: Ucok et al found no difference in respiratory function when comparing 31 women with PCOS vs age- and BMI-matched controls (34).

In the present study, cases and controls were not matched for BMI, since an increased BMI is part of the PCOS feature in most women with PCOS (35). Additionally, excess weight exacerbates the overall expression of the syndrome (36). Although the incidence of asthma in women increases with BMI (37), adjusting for BMI would indirectly also adjust for the impact of PCOS. Subgroup analysis of nonobese participants showed a trend towards reduced respiratory function in PCOS in our population. Therefore, it is less likely that the increased prevalence of self-reported asthma in PCOS can be explained by obesity alone. Obesity itself could possibly predispose to a false diagnosis of asthma; however, a population-based study found similar confirmation rates of asthma in obese (68%) versus nonobese (71%) individuals (38). The higher rate of reported asthma in PCOS is therefore less likely to be due to a diagnostic bias in individuals who are obese.

Fide et al found higher prevalence of “adult onset asthma,” defined as asthma diagnosed 1 or more years after age at menarche, among women who reported early menarche than in women who did not report early menarche (39). Patient-recall of onset of asthma over a decade later has been found reliable in a Swedish cohort (40). A longitudinal cohort study reported more wheezing at both 11 and 13 years of age in girls with excess weight at 11 years of age, and the highest prevalence of wheezing was described in overweight and obese girls with concurrent reported start of puberty at younger than 11 years of age (41). Although it is important to bear in mind that wheezing is not synonymous with asthma, wheezing at the end of the first decade of life has a persistent nature, and the prevalence of wheezing increases with the prevalence of asthma in girls (42). Ibáñez et al described that premature adrenarche predisposed to features of PCOS in adolescent years, which were improved by metformin, indicating a prominent role of insulin resistance during these transitional years (43). In our population, the prevalence of reported menarche at younger than 13 years was higher in the PCOS group, and the PCOS group had a higher incidence of reported asthma in the peripubertal period (11–15 years) compared with the control group (Fig. 3). In a Mendelian randomization study, adult FVC, but not FEV1/FVC, increased with increasing age at menarche; the authors suggested that early puberty initiated premature flattening of lung development (44). Childhood obesity, insulin resistance, and early onset of puberty are clinically related features that seem to underpin the development of impaired lung function, as well as a restrictive pattern, in PCOS. The present study demonstrated impaired respiratory function by reduced FEV1 % predicted and FVC % predicted values in women with PCOS compared with controls, both for self-reported asthma and nonasthma (Fig. 2). FEV1 % predicted was not different in asthmatic versus nonasthmatic women with PCOS, which raises the question regarding the extent to which the increased prevalence of reported asthma in PCOS—both previously reported and in the current study—represents true asthma. A genetic component of PCOS development has been suggested and a recent study showed correlation between genetic susceptibility locus of PCOS and a broad range of metabolic disorders (45). Impaired lung development in PCOS, due to exposure to diabetes or other adverse factors in utero, could be an alternative explanation to our findings (46, 47).

The higher prevalence of reported asthma was not accompanied by more reported hay fever or nasal allergies in the PCOS group. This is in line with Chen et al who described a higher incidence of asthma in nonallergic obese women than in allergic obese women (48). Obesity causes a proinflammatory environment that might worsen respiratory function by inflammatory cell recruitment, altered airway immune response, increased smooth muscle constriction, and/or increased airway hyperresponsiveness (49). Adipose tissue also function as an endocrine and paracrine organ which releases cytokines and bioactive mediators (50) (Fig. 4). A possible pathogenic mechanism for asthma in obesity may be a direct effect on airway epithelial cells of pro-inflammatory adipokines. Sideleva et al found a correlation of visceral fat leptin expression with airway reactivity (51). Disturbances in adipokine levels have been reported in PCOS (52). An association between PCOS and both chronic low-grade inflammation, reflected by minor but significant increase of mediators of inflammation such as high-sensitivity C-reactive protein and migration inhibitor factor, and autoimmunity has been suggested (53–55). There are also indications of autoimmune involvement in asthma, both by associations of circulating autoantibodies and asthma and by common gene polymorphisms and susceptibility genes associated with both asthma and more established autoimmune diseases such as systemic lupus erythematosus and Crohn disease. This suggests linkage of T helper type 1 lymphocytes (Th1)- and Th2-immune mechanisms, rather than the traditional Th1/Th2 paradigm, in which the adaptive immune system is activated by infectious agents and/or autoantibodies triggering off tissue damage and chronic inflammation (56). Our findings indicate that nonatopic adult-onset type of asthma, with more persistent symptoms, dominates in PCOS and we suggest that dysfunctional adipose tissue and chronic low-grade inflammation and/or autoimmunity may be important predisposing factors in this association (37, 57–60).

Figure 4.

Respiratory impairment of a combined obstructive-restrictive pattern in PCOS may be caused by direct or more indirect pathways.

Women with PCOS had higher education than the controls in our study; this is possibly due to demographic difference in the populations studied. The original pilot and PregMet studies recruited predominantly from urban areas, while HUNT-controls were recruited from both urban and rural areas (29). A longitudinal cohort study from New Zealand found no association of socioeconomic status or educational achievement with current asthma, DDA, current wheeze, time off work due to asthma, or bronchodilator response (61). Surveillance bias could not potentiate our finding of more asthma in the PCOS group, as both women with PCOS and controls have been approached repeatedly. The higher level of reported physical activity in PCOS might be due to reporting bias, (ie, over-reporting of activity in PCOS), as respondents know they are “at risk” and more prone to metabolic dysfunction (50).

We found more self-reported asthma and impaired respiratory function in women with PCOS compared with controls. The respiratory impairment showed a combined obstructive-restrictive pattern in PCOS, not previously reported. Further studies are required to characterize possible mechanisms of inferior respiratory health in PCOS.

Abbreviations

Abbreviations
 
• BMI

  body mass index

 
• DDA

  doctor-diagnosed asthma

 
• HOMA-IR

  homeostatic model assessment for insulin resistance

 
• FEV1

  forced expiratory volume in the first second of expiration

 
• FVC

  forced vital capacity

 
• LC-MS/MS

  liquid chromatography-tandem mass spectrometry

 
• OR

  odds ratio

 
• PCOS

  polycystic ovary syndrome

 
• SHBG

  sex hormone-binding globulin

Acknowledgments

We thank The HUNT Study for providing the data from the control-group. The HUNT Study is a collaboration between the HUNT Research Centre (Faculty of Medicine and Health Sciences, Norwegian University of Science and Technology), Nord‐Trøndelag County Council, Central Norway Health Authority, and the Norwegian Institute of Public Health. We are grateful to all participants for taking part in this study.

Financial Support: The Research Council of Norway (NFR), Novo Nordisk Foundation Norway, Felles Forskningsutvalg (Norwegian University of Science and Technology, Faculty of Medicine and Health Sciences/St.Olavs Hospital trust)

Additional Information

Disclosure Summary: The authors have nothing to disclose.

Data availability: Restrictions apply to the availability of data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.