In this study we tested whether a 6-month treatment with the ethinylestradiol/norelgestromin contraceptive patch induced changes in body composition, particularly in body fluid volume. After a treatment of 6 months we documented a small but statistically significant increase in body weight, 0.64 kg (1.1%), that has been due to an increase in soft tissue hydration of similar amount (TBW 0.51 L, 1.7%). This quantity of fluid is within the physiological range of interstitial gel’s hydration fluctuations [18, 19]. In fact, we interpreted this small increase in soft tissue hydration as a physiological change in the interstitial gel fluid which may represent an adaptation to the estrogen component of the patch. As gel fluids are not free to move inside vessels, it does not increase the blood volume [18, 19]. Consistently, we did not document any effect on blood pressure.
Therefore, the changes in body weight occurred in our study are clinically not relevant. None of our women dropped out for weight gain, fluid retention or perceptions of them. In this study, women's acceptability and satisfaction with the contraceptive patch were high, their compliance remarkably good and, after 6 months, the majority of subjects indicated that they would continue using the patch.
According to Guyton's theory , fluid overload is detectable as apparent edema when interstitial pressure becomes positive due to an increase of interstitial fluid volume above 30% (meaning increase > 4–5 kg body weight, or > 12% TBW). In a normal subject (40 liters of TBW), the interstitial fluid volume is 15 liters in a gel form and with a negative interstitial pressure of – 2 mmHg, the blood volume is 5 liters, and tissue impedance is normal. When the interstitial fluid pressure rises above zero most of the extra fluid is free fluid allowing the appearance of pitting edema and shortening the impedance vector due to a decrease of both vector components. Hence, vector migration on the R-Xc plane can be used in monitoring tissue hydration before the appearance of clinical signs of fluid overload [16, 19].
Also a recent review on transdermal hormonal contraception concludes that patch users generally experience minimal, clinically not relevant, changes in body weight . In a recently published meta-analysis on 47 articles  there was no evidence supporting a causal association between combined oral contraceptive or a combined skin patch and weight gain. Interestingly, in 15 studies that used ethinylestradiol 20 μg in one arm of treatment, 4 considered weight change at 3, 4, or 6 months and found a range from -0.11 to 0.88 kg (compared to our 0.64 kg) [20–23]. Three studies considered mean body mass percentage change (1.1% as in our population). Other studies evaluated the odds ratio of a relevant weight gain, that was set to > 2 kg of weight or >5%. Analyzing the daily weights of 128 women in treatment with oral contraceptives, Rosenberg found that there were minor fluctuations, clinically not relevant, of body weight during each cycles, confirming that weight gain is a myth or a misperception .
We used BIA methods to establish the nature of the body weight change. Clinical validation studies established that a change in the body weight due to a fat mass change in the order of 10 kg does not displace the impedance vector. In contrast, a change in the body weight due to a fluid volume change in the order of 0.5 L is detected as a vector displacement parallel to the major axis of vector's tolerance and confidence ellipses (i.e. change in both R/H and Xc/H components) [2, 16, 25]. Conventional BIA equations are functions not only of R but also of the body weight. Any body weight change will yield a change in TBW and in both fat and fat-free mass due to the assumption of fixed hydration of soft tissues . With Vector BIA we documented the same pattern of increased hydration of soft tissues as found by Lukaski in pregnancy in a larger scale of hydration . With conventional BIA we estimated an increase in TBW that was in the expected range for normal tissue hydration (TBW = 73% of 0.64 kg soft tissue, i.e. 0.47 L). After 1 month with transdermal patch there was no significant difference in changes in body weight, impedance vector, nor TBW among previous contraception methods (Table 2).
Although we are more confident in Vector BIA as a tool for body composition in any clinical condition [16, 19], we estimated TBW for the sake of comparison with the scanty literature, where impedance values are not reported but comparable TBW values were found [26, 27]. We found conflicting indications from TBW vs Vector BIA, probably due to the high standard error of BIA prediction equation estimates [4, 15]. For instance after 3 months, the estimated TBW increase (0.49 L) was greater than the body weight increase (0.42 kg); at T0 there was a significant difference in TBW among the three previous contraception methods, although body weight and impedance vector did not differ among the three methods (Fig. 2). This may occur because the prediction error of BIA equations is the sum of five errors, namely the impedance measurement error, the regression error against the reference method, the intrinsic error of the reference method, the electric-volume model error, and the biological variability among subjects. Vector BIA only needs to take care of the measurement error and of the biological variability of subjects. Hence no change in impedance means no change in hydration independent of the body weight change [16, 19].
However, estimates of TBW and of derived compartments (fat and fat-free mass) with BIA prediction equations are not useful in the individual subject due to the high prediction error [4, 5, 15]. Even weaker is the validity of the estimation of intra- and extracellular fluid with multifrequency BIA [16, 28, 29] due to tissue anisotropy (part of the current flow is intracellular at low frequencies). For these reasons we used the standard, 50 kHz frequency current (which has the best signal-to-noise ratio) and the Vector BIA which is very sensitive to biological variation of measurements without the need of regression equations [4, 16].
The sensitivity of Vector BIA in detecting changes in soft tissue hydration of less than 1 L (0.64 kg of body weight) could help clinicians in routine, objective monitoring of the patients’ body fluid volume, particularly with patients who have the perception of increased body fluid volume of patients, particularly of those with the perception of increased body fluid volume.
Our findings were documented in a "real world" population by using broad criteria for age (18 to 45 yr) and BMI (16 to 31 kg/m2). Validity of results can therefore be extended to the routine clinical setting, while this is not possible for other studies on extremely selected subjects (e.g. age 18 to 38 yr, and BMI 21 to 25 kg/m2) . Finally, even if our women had been monitored only for 6 months, we are quite confident that we would have obtained more or less the same results with a longer study period. As for body weight, in other published studies the patch was used for up to 13 cycles showing no significant signs of weight gain. As for body composition, other published trials evaluating the effect of oral contraceptives on body water and body fat lasted for a maximum of 6 months [14, 26, 27, 30, 31].
An important limitation of the study is the lack of a control group. However, a randomised controlled trial comparing a combined contraceptive method with a placebo or non-hormonal method for contraception raises ethical issues. As a rough surrogate we classified subjects according to their contraception method used in the three months before study entry (no method versus oral contraceptives versus barrier method) and we compared changes in body composition at entry and after 1 month of treatment. With Vector BIA we proved that there was neither a baseline difference nor a differential change in body composition associated with contraception methods used in the three months before.
In any case, present findings will prompt a comparative, or, if ethically acceptable, a placebo/non-hormonal controlled trial with a longer follow up period.