In the present study, the administration of 1 or 2 grams of essential fatty acids to patients with PMS resulted in a significant decrease in symptom scores, as evaluated using the PRISM calendar. The three groups analyzed were well-balanced with respect to the age, ethnicity, marital status and schooling of the patients, confirming the validity of the randomization procedure.
Various diagnostic scales are available; however, the PRISM calendar was selected as being one of the best known and most widely used in clinical and epidemiological studies on PMS . It consists of 23 questions on symptoms and their intensity during the menstrual cycle and is compatible with the criteria defined in the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR). This self-applicable, relatively simple questionnaire is adequate for evaluating large populations within a short period of time and allows quantification of the symptoms reported by the patient and a comparative analysis between individuals.
One of the strong points of the present study lies in the rigorous inclusion criteria. If on the one hand these stringent criteria made the admission of patients to the study more difficult, on the other hand they contributed by minimizing potential biases such as contraceptive use, obesity, organic diseases or psychological disorders that could have affected symptoms. The result was a rigorously selected sample population that was highly motivated to participate in the study, so increasing the internal validity of the study. This can be clearly seen from the fact that none of the participants missed a visit or abandoned treatment during the eight months of follow-up. Only four patients were excluded from the analysis, one because she had hyperprolactinemia, which was detected following admission to the study but before initiating the study medication, a second because of a thyroid disorder and the other two because they were found to be in use of antidepressants that could have hampered analysis of the results.
The use of the PRISM calendar in the first two months of follow-up served to identify women with PMS and differentiate them from patients with psychological disorders, since in the latter group symptoms do not improve at any time during the menstrual cycle. In the period immediately preceding treatment, statistically significant differences were found in the overall PRISM scores in the women in the three study groups when the scores for the follicular phase of the cycle were compared with the scores for the luteal phase, showing that a significant increase in symptomatology did occur within the same month, thus characterizing PMS.
A decrease in PRISM scores in both the follicular and luteal phases was observed in all three groups, reflecting an improvement in symptomatology. However, there was a significant difference in the magnitude of the reduction between the groups using the medication and the placebo group. When the absolute difference between the symptom score in the follicular and luteal phases of each group was analyzed throughout the treatment period, the groups were found to be paired with respect to the difference in score points. Moreover, this absolute difference, which reflects the intensity of PMS symptoms within one single month, decreased gradually in all three groups analyzed. However, there was a statistically significant difference between the groups using either 1 or 2 grams of the medication and the placebo group. This difference was already evident at three months and became even more apparent after six months of treatment. After only three months of treatment, the effect of the medication on PMS symptoms was already significant, whereas in the patients in the placebo group this improvement was less noticeable. Furthermore, after three months of treatment, clinical improvement was bigger in the case of the women in the 2-gram group compared to those in the 1-gram group, showing that the higher dose of the essential fatty acids contained in these pharmacological preparations resulted in a higher reduction in symptoms.
Analysis of the absolute and relative differences between the overall symptom score in the follicular and luteal phases of the cycle throughout the treatment period in the three groups evaluated showed that scores of symptoms diminished significantly, both in the follicular phase and in the luteal phase in groups A (1 gram of medication) and B (2 grams of medication), while the decrease in group C (the placebo group) was more discrete. However, this decrease in scores of symptoms in the placebo group after six months of treatment was no longer statistically significant. These data support the hypothesis that this medication effectively reduces PMS symptoms [12, 16, 23].
The initial clinical improvement observed in patients in group C (placebo) was probably due to the "placebo effect", an important factor that is widely recognized in the literature and describes a phenomenon that occurs when a clinical improvement is found in an effect under analysis in a person or group in which the treatment given was inert . When dealing with PMS patients, these psychological effects are even more important than in other situations, since, within the physiological and pathological bases of this syndrome, the emotional factor is of utmost importance. Patients with PMS are generally vulnerable and distressed by their cyclic symptoms, which may be debilitating. Psychosocial management is, therefore, essential and should involve the interaction and education of family members, as well as lifestyle changes and medication. Data from the literature show that an improvement of as much as 50% in symptoms is found in up to 20% of patients submitted to placebo treatment in PMS studies .
Many PMS symptoms are similar to the effect produced by an injection of prolactin [12, 25]. Some women with PMS have high prolactin levels; however, levels are normal in the vast majority of patients. Women with PMS may be abnormally sensitive to normal amounts of prolactin  and this phenomenon may be associated with low PGE1 levels.
This could be a consequence of the fact that PGE1 acts on almost all organs of the body. It has a diuretic effect by promoting a reduction in angiotensin II. Fatty acids from food intake alter hormone and neuropeptide levels such as norepinephrine, dopamine and serotonin. Fatty acids also affect receptors for hormones and neuropeptides  and, through PGE1, affect tissue sensitivity to prolactin. There is evidence that prostaglandin E1 is able to attenuate the biological effects of prolactin and that, in the absence of prostaglandin E1, the effects of prolactin are exacerbated .
The results of this study confirm the findings of other authors who have recommended polyunsaturated fatty acids as a therapeutic option for patients with PMS [12, 20, 22, 23]. Many studies have shown the efficacy of nutrients on PMS symptoms. Most report an improvement, mainly in emotional symptoms, with the use of pyridoxine (vitamin B6) . Ascorbic acid and niacin have also been mentioned. Pyridoxine deficiency has already been suggested as a cause of PMS [27, 28]. Magnesium hypoactivity has also been associated with different pathological states such as PMS, since magnesium levels are closely related to the activity and secretion of gonadal hormones and this may contribute towards the genesis of this condition [29, 30]. Nonetheless, the clinical success obtained with some of these nutrients may be partially related to their effects on essential fatty acid metabolism and PGE1 production, since the delta-6 desaturase enzyme requires the presence of zinc, magnesium and insulin to exert its effect, while the formation of gamma-linolenic and dihomo-gamma-linolenic acids requires pyridoxine as a cofactor. On the other hand, COX-1 requires the presence of niacin, vitamin C and zinc.
Currently, serotonin reuptake inhibitors (5-HT) are gaining popularity for the treatment of PMS, since studies show that a deficiency of this substance may be involved in the etiology of the condition . Therefore, serotonergic antidepressants such as sertraline, fluoxetine, citalopram and clomipramine have been shown to be effective for intermittent use in the luteal phase of the menstrual cycle , mainly in patients with PMDD, resulting in a reduction in emotional and physical symptoms. Studies have shown no differences on the effects of this medication in the treatment of PMS, and particularly PMDD, when use is continuous or restricted to the luteal phase; therefore, intermittent use is recommended [33–35].
To evaluate whether essential fatty acids would alter prolactin levels by increasing PGE1 levels, this hormone was measured during the luteal phase at the beginning and at the end of treatment. When prolactin levels were compared in the three groups evaluated over the six months of treatment, no statistically significant differences were found between baseline values and levels measured at the end of the treatment period, showing that the medication had no direct effect on prolactin. This reinforces the hypothesis that the improvement in symptoms is probably due to alterations in tissue sensitivity to this substance .
One concern when administering essential fatty acids as a dietary supplement is their effect on lipid indexes. To evaluate this effect, total cholesterol was measured prior to and following treatment. No statistically significant difference was found between the groups, or between the evaluation moments during treatment, showing that the administration of a dietary supplement of essential fatty acids did not result in any changes in total cholesterol in the patients evaluated.
These findings confirm results published in the literature showing that no hormonal or biochemical changes occurred with the use of essential fatty acids in patients with PMS .
Few adverse events were recorded and these were mild, insignificant and did not appear to be directly related to the medication. The two patients in the placebo group who suffered episodes of diarrhea may be excessively sensitive to mineral oil, since the dose given was too low to act as a laxative. Nevertheless, these patients later reported having had no further episodes of these symptoms.