The research was carried out in the period from 2015 to 2017 among 1257 female university students aged 19 to 25 years. The study protocol was approved by the Bioethics Committee of the Jagiellonian University (opinion no.122.612047.2016). Data were collected following the ethical principles as stated in the Declaration of Helsinki. Informed consent was obtained from all individual participants included in the study.
Age at menarche was assessed based on the date provided by the subjects. Students were asked to provide the date of their first menstruation with the accuracy within one month or a quarter. The age at menarche was calculated from the date of menarche and the date of birth. In cases when, instead of an exact date, participants reported only month and year of menarche, the 15th day of an indicated month was used for calculation, and when the date of menarche was reported approximately in the range of 2–3 months, the midpoint of the reported period was used for calculation.
The level of each analyzed environmental pollution: particular mater (PM10, PM2.5), sulphur dioxide (SO2), nitric oxide (NO) and benzene (C6H6) in the town of residence during childhood and adolescence was be determined on the basis of the data made available by the Chief Inspectorate for Environmental Protection. Previous place of residence was divided into 3 categories, taking into account the air quality class to which the area was classified for most of the time (at least within 10 years). Class 1 comprised zones where annual pollutant values and the number of days per year with exceedances were below the allowable limit; Class 2 - zones with annual values below the permissible limit, but with the number of days of exceeding the norm above the limit, Class 3 included zones above the limit. Based on all the level of each pollutant a complex air pollution index was established. To assess the joint effect of air pollutants, principal component analyses were used to group pollutants (PM10, PM2.5 SO2, NO, and C6H6) to represent a source-related mixture. Since the first component was relatively high, scores for the single component were used as an indicator of air quality. Tertiles of the component were used to identify three air quality classes: good, moderate or unhealthy.
Socio-economic status was included in the analysis as covariates and it was determined on the basis of variables considered reliable indicators of living conditions and lifestyle in Poland and Europe: degree of urbanization of the place of residence, father’s and mother’s education, number of siblings, and financial situation. The following categories were created: for the place of residence in childhood and adolescence: village, city up to 100,000 inhabitants, city above 100,000 inhabitants; for mother’s and father’s education: vocational, secondary, higher; for number of siblings: none, one, two, three or more. Financial situation during childhood and adolescence was assessed based upon subject’s responses to a survey question: ‘Do you consider the economic situation in your home as: below average and bad, average, good, very good, changeable and/or hard to specify’. As there was only one response in the last category, it has been omitted. Based on all the above data, a SES evaluation index was established and subjects were qualified as belonging to families of low, average, or high SES. The division was introduced on the basis of the value of the first component obtained in the principal components analysis (PCA).
The study design is fully described elsewhere [5].
Statistical Methods
Statistical analyses were performed with the use of statistical software Statistica 13.0 (by StatSoft Polska). This manuscript is a continuation of our research analyzing the impact of air pollution on biological development therefore, we used the same statistical methods as previously [5]. The Shapiro-Wilk test was applied for stature distribution normality assessment, and the Levene’s test for assessing the equality of variance of the analysed data. Air pollutions were analysed as categorical variables. In statistical analysis the generalized linear model (GLM) was applied. Analyses were conducted in three steps. Model 1 included the level of each analysed air pollutant: particulate matter (PM10, PM2.5), sulphur dioxide (SO2), nitric oxide (NO) and benzene (C6H6). Model 2 was adjusted for the degree of urbanization of the place of residence. In case of subject who moved during childhood or adolescence, the degree of urban development was calculated as the average category of all residences of the subject, weighted by the length of each residential period. Model 3 included as covariates all socio-economic variables: degree of urbanization of the place of residence, father’s and mother’s education, number of siblings, and financial situation. All adjustment covariates were selected a priori [5].
To assess the effect of multiple-pollutant exposure on the age at menarche two-way analysis of variance and logistic regression were applied. Significance in all statistical tests was set at the level of at least p<0.05 [5].