Variability of thyroid blood flow Doppler parameters in healthy women

Abstract

The purpose of this study was to estimate variability of flow Doppler parameters in the superior thyroid artery (STHA) during the menstrual cycle in young women and to explore the influence of endogenous 17-b-estradiol (E2) and progesterone (PRG) on the velocity waveform. The plasma concentration of these hormones was correlated with flow velocities, pulsatility index (PUI), resistance index (RI) and acceleration index (accI) and time (accT), which were measured with color-coded duplex sonography 8 times during the cycle in 14 healthy women (age range: 23 to 25 years). Coefficient of variation (CV), interclass correlation (ICC), repeatability (repC) and pooled Pearson correlation (r) coefficients were used to estimate the variability of the parameters. The highest variability was found for accI and accT: CV = 48% and 31%; ICC = 0.51 and 0.45; repC = 2.8 and 95; r = 0.37 and 0.4, respectively. The CV for flow velocities varied from 25% to 26%, ICC from 0.53 to 0.56, repC from 8 to 17 and r has a value of 0.46. The respective values for RI and PUI were: 11%, 18%; 0.48, 0.55; 0.15, 0.48; and 0.46, 0.48. The diastolic blood pressure decreased significantly by 7 mmHg (p < 0.01) in the luteal phase, whereas other physiological variables were stable during the cycle. Although the fluctuations of the flow parameters during the cycle were not statistically significant, a weak linear correlation between flow velocities and concentration of E2 was found; for mean velocity r = 0.16, p < 0.05. Impedance indices showed an increasing trend in the luteal phase, along with increase of the pulse pressure index (PPI). The results showed that variability of the flow parameters in the STHA is substantial and that higher flow velocities are associated with increase of plasma concentration of 17-b-estradiol during the menstrual cycle in young women. (E-mail: [email protected])

Introduction

The increased blood flow through the thyroid gland has been observed in several diseases, such as Graves’ disease Ralls et al 1999, Saleh et al 2002, Huang et al 2003, Hashimoto’s thyroiditis Vitti et al 1995, Caruso et al 2000, amiodarone-induced thyrotoxicosis Bogazzi et al 1997, Bogazzi et al 2001 and pituary gland tumors producing thyroid-stimulating hormone (TSH) (Bogazzi et al. 1999). This led to a premise that the level of blood flow through the gland can be a useful index of thyroid function Woodcock et al 1985, Hodgson et al 1988, disease activity Castagnone et al 1996, Baldini et al 1997, Saleh et al 2001 and stimulation of the TSH-receptor (Bogazzi et al. 1999). However, little is known about the range of variability of the blood flow in healthy subjects.

Color Doppler sonography is widely used to image blood flow through the microvasculature of the thyroid gland, but assessment of blood flow on the basis of color flow mapping is subjective. Color-coded duplex sonography allows measurement of blood flow velocity in major arteries supplying the thyroid gland (Fobbe 1995). Blood flow velocities and other Doppler indices might be more objective parameters for evaluating thyroid dysfunction (Castagnone et al. 1996). The flow through the thyroid arteries is modified by many frequently uncontrolled factors that can limit the diagnostic accuracy of blood flow Doppler parameters Schweiger et al 1996, Chan et al 1998. Therefore, assessment of between-subjects and within-subject variability of the parameters is of particular importance in differentiating “normal” from “abnormal” flow patterns in follow-up studies.

Recently, Chan et al. (1998) have shown that the pulsatility index (PUI) in the superior thyroid artery varies significantly throughout a menstrual cycle in healthy women. Although plasma concentrations of estrogen and progesterone in their subjects were not measured, they suggested an influence of estrogen on thyroid hemodynamics. Therefore, it can be hypothesized that substantial natural fluctuation of plasma concentration of 17-b-estradiol (E2) and progesterone (PRG) during the menstrual cycle will contribute to variability of flow Doppler parameters.

The purpose of this study was: 1. to estimate variability of blood flow Doppler parameters in the superior thyroid artery during the menstrual cycle in young, healthy women, and 2. to explore whether or not the parameters are influenced by fluctuations of plasma concentration of endogenous E2 and PRG during the cycle.

Institutional Ethical Committee approval was obtained for this prospective study and each volunteer gave written informed consent.

A total of 16 women medical students (age range: 23 to 25 years, mean weight 56.4 ± 9.7 kg) who met strict inclusion criteria initially enrolled into this prospective study. In particular, none of them was a smoker, none was taking any medication at the time of the investigation and none had ever used oral contraceptives or hormones. All had six self-reported regular consecutive menstrual cycles of approximately 28 days (range 27 to 32 days) before the investigation.

All subjects underwent a screening clinical examination performed by a physician. Physical examination, blood cell count, liver function tests and the concentrations of electrolyte, blood glucose, cholesterol and triglycerides were within normal limits. Subjects with enlargement of the thyroid gland, abnormal echotexture or echogenicity, as determined with grey-scale sonography in the beginning of the follicular phase, were not included. Serum concentrations of thyroid-stimulating hormone (TSH) in our subjects were within normal range (mean 1.407 ± 0.794 mIU/L, range 0.45 to 3.52).

Of 16 subjects who entered the program, 2 were excluded at the final stage of the study, 1 because of an unovulatory cycle and the other because of an episode of fever during the study period. Thus, 14 women were ultimately included in the study.

Every participant was evaluated at least 10 times: during menses (cycle days 1, 2, 3), during the follicular phase (cycle days 6, 12, 13, 14) and during the luteal phase (cycle days 15, 20, 24). Two initial examinations, taken on days 1 and 2 of the cycle, were not included in analysis because they were designed to minimize the effects of anxiety in the subject. Ovulation was determined by direct sonographic follow-up of the follicle and with measurements of plasma E2 and PRG concentrations. The cycles were counted from the first day of menses and were standardized to a 28-day period. Ovulation was taken as occurring between days 14 and 15.

On the days of testing, subjects reported to the laboratory after having fasted for 12 h, and had abstained from vigorous exercise, alcohol intake and caffeine-containing beverages for at least 24 h before the study. All examinations were performed between 6 and 8 a.m. to minimize the effect of circadian rhythms on the cardiovascular system. The study was carried out in a quiet room, with subjects lying in the supine position, after a 15-min rest period. Brachial cuff systolic (SBP), diastolic (DBP) blood pressure and heart rate (HR) were measured before and after having completed all sonographic studies. The mean value from the two measurements was taken for further analysis. A pulse pressure index (PPI) was computed using the formula [SBP-DBP]/SBP. A sample of blood was collected after the sonographic study to determine hematocrit, hemoglobin and plasma concentrations of E2 and PRG. The concentration of these hormones was measured immediately by an automated chemiluminescence system (ACS 180:PLUS immunoassay; Bayer, Leverkusen, Germany). The measurements were carried out with errors of less than 12% and 9%, respectively, expressed as CV (provided by the manufacturer).

We used a 7.5-MHz linear transducer (Siemens Elegra, Siemens, Erlangen, Germany) for B-mode and color-coded duplex imaging. Doppler settings for color mode were: Doppler detection frequency, 5.1 MHz; pulse repetition frequency (PRF), 2 kHz; and, for duplex mode: Doppler frequency, 5.5 MHz; PRF, 4 kHz; sample width, 2 mm; and wall filter 50 Hz. All examinations were performed by one investigator with 8 years of experience in Doppler sonography. Scanning was performed with the patient in the supine position with neck hyperextended and rotated away from the operator. Using an anterolateral approach, a longitudinal view of the right common carotid artery was obtained and the superior thyroid artery (STHA) was localized by slowly moving the transducer medially from the common carotid artery. The STHA was identified as having blood flow in the opposite direction from that of the common carotid artery and having high diastolic flow showing on the Doppler spectrum (Fig. 1). The transducer was then manipulated to align with the longitudinal axis of the STHA. The measurements were made at the superior pole of the thyroid gland. Only the right STHA was studied, to obtain more reproducible results. The angle of insonation was always less than 60°, to maintain higher accuracy in measurements. The peak systolic (v_ps), mean (v_m) and end-diastolic (v_ed) velocities were obtained by automatically outlining the maximum frequency envelope of the Doppler waveform over completed cycles. Pulsatility (PUI) and resistance (RI) indices were calculated using the formulas provided by Gosling and King (1974) and Pourcelot (1974), respectively. The acceleration time (accT) and acceleration index (accI) were also measured from the Doppler spectrum. The accT was defined as the time between the beginning of the systolic upstroke and the first systolic peak. The accI was defined as the gradient of this upstroke (Stavros et al. 1992). The values of all Doppler parameters in each subject were also standardized by relating them to the base value of the respective parameter obtained during the initial examination in this subject taken at day 3, and were given as percentages to better demonstrate the trends of the Doppler parameters during the menstrual cycle.

The data were analyzed using a statistical software (SYSTAT® for Windows, SYSTAT, Evanston, IL). For all measured Doppler parameters, the mean, range, median and SD (pooled) across subjects were calculated during the cycle. The normal distribution of the measurements was verified by the normal plot method.

Several statistical measures were used to estimate the variability of blood flow Doppler parameters during the menstrual cycle.

First, to estimate the within-subject and between-subjects variance components for each Doppler parameter, a one-way random-effects ANOVA was employed. Computations were based on both original and transformed data to a logarithmic scale. Second, to estimate reproducibility of replicate measurements from the same subject, the interclass correlation coefficient (ICC) was computed for eight measurements obtained from 8 days of the cycle Rosner 1995, Giraudeau and Mary 2001. The coefficient ICC is the ratio of the between-persons variance divided by the sum of the between-persons and the within-person variance. The absolute values of ICC vary between 0 and 1, with a value of 0 indicating no reproducibility at all and a value of 1 indicating perfect reproducibility. For a measurement to be useful, it has been recommended that ICC should be at least 0.60 (Giraudeau and Mary 2001). Third, the coefficient of variation (CV), as a measure of reproducibility, was estimated by computing the square root of within-subject mean square obtained from one-way random-effects ANOVA based on data transformed to logarithmic scale. In general, CV < 20% is desirable and CV > 30% is undesirable (Rosner 1995). Fourth, the repeatability coefficient (repC), which represents the value below which the difference between the repeated measurements will lie with a probability of 95%, was computed as 1.96 × square root of 2 s², where s² is within-subject variance. Fifth and last, the pooled Pearson correlation coefficient (P_r) was computed by averaging the Pearson correlation coefficients (r) from each pair of measurements performed on any two separate days.

Repeated measures ANOVA was used to test the null hypothesis stating that standardized blood flow Doppler parameters, blood pressure and heart rate did not change during the menstrual cycle. In case of significant differences, however, the paired t-test was used with the optional p value adjustments (Dunn-Sidak) for multiple tests. Multiple regression analyses were used to quantify relationships between SBP, DBP, HR, PPI, E2, PRG, angle of insonation and each blood flow Doppler parameter. Interrelations of SBP, DBP, HR and PPI were taken into account. Levels of probability less then 0.05 were considered to be statistically significant.