Finally, we collected 2833 papers for bibliographic records after evaluating 3,852 original results from the search for PMS/PMDD-related papers published in the Web of Science from 1945 to 2018 on the basis of the inclusion and exclusion criteria.
Annual publication outputs
As shown in Figure 2, we counted the number of publications for each year. Overall, there was an upward trend in publications from 1950 to 2008, but there were fluctuations in some years. However, we can see that the number of papers published in a single year reached a maximum of 124 in 2008. Then, there was a slow downward trend from 2008 to 2018. Moreover, we can see that the trend in outputs was not stable; since 2000, there were fluctuations in 2001, 2007, 2009, 2012, 2016 and 2017. This finding may be caused by the publication cycle. On the one hand, an upward trend indicates that relevant research is becoming a hotspot, but a download trend indicates that relevant research attention tends to be flat.
Due to restrictions on the year (since 1945), the literature was retrieved from WOS. We did not find Frank’s original paper[8], which is regarded as the opening work of PMS/PMDD research. Publications about PMS/PMDD could be traced back to 1950. In 1950, three articles were published: “Nephrotic syndrome with exaggerated premenstrual water and salt retention”, written by RW Lippman[17]; “Premenstrual tension”, written by JH Morton[18], who worked at the New York Medical College; and “Premenstrual tension treated with vitamin A”, written by J Argonz and C Abinzano [19]. Among these three articles, JH Morton’s article had the biggest impact, with 116 citations. More in-depth research findings showed that Morton published 80 articles from 1945 to 2010, mainly on surgery (35), general internal medicine (21) and obstetrics gynecology (15). Research institutions were mainly in the USA (University of Rochester).
Distribution of journals and co-cited journals
In total, 295 academic journals have published articles on PMS/PMDD. According to the Journal Citation Reports (JCR) 2017 standards, the top 15 journals contributing to PMS/PMDD are shown in Table 2. Psychneuroendocrinology (impact factor (IF) 2017 = 4.731) published the most papers (79 publications, 2.789%), followed by the Journal of Psychosomatic Obstetrics and Gynecology (IF 2017 = 1.900, 61 publications; 2.153%), the American Journal of Obstetrics and Gynecology (IF 2017= 5.732, 50 publications; 1.765%) and Gynecological Endocrinology (IF 2017 = 1.453, 49 publications; 1.730%).
In addition, Biological Psychiatry (IF 2018 = 11.982) had an IF higher than 10, and Gynecology (IF 2018 = 5.732) had an IF between 10 and five. Five journals, Obstetrics and Gynecology (IF 2018=4.982), Psychoneuroendocrinology (IF 2018=4.731), the Journal of Clinical Psychiatry (IF 2018=4.247), Psychosomatic Medicine (IF 2018=3.810), and the Journal of Affective Disorders (IF 2018 = 3.786), had an IF between five and three.
We used CiteSpace’s dual-map overlay function to construct a citation dual-map to visualize a more comprehensive view of the citation state of PMS/PMDD. Dual-map overlays could show the interactions of journals. As shown in Figure 3, the left and right sides correspond to the citing and cited journal maps, respectively. The labels represent the disciplines covered by the journal. The lines on the map starting from the left and ending on the right represent the citation links. There were eight citation paths. The upward yellow path shows that papers published in immunology/biology journals mostly cited journals in the area of biology/genetics; the downward yellow path shows that papers published in immunology/biology journals mostly cited journals in education/social areas; the upward green path shows that papers published in medicine/medical/clinical journals mostly cited journals in the area of molecular biology/biology/genetics; the middle green path shows that papers published in medical/clinical journals partially cited journals in the health/nursing area; and the bottom green path shows that papers published in medicine/medical/clinical journals partially cited journals in the psychology/education/social area. The upward blue path shows that papers published in psychology/education/health journals mostly cited journals in the molecular biology/biology/genetics area; the middle blue path shows that papers published in psychology/education/health journals partially cited journals in the health/nursing/medicine area; and the bottom blue path shows that papers published in psychology/education/health journals partially cited journals in the psychology/education/social area (Figure 3)
Distribution of countries/regions and institutes
The 2,833 publications on PMS/PMDD were contributed by 76 countries/regions. There was extensive collaboration between countries/regions (Figure 4). As shown in Figure 4, colors showed different research directions. The larger nodes represented the more influential countries in this field. In relation to the top 10 countries that contributed PMS/PMDD research, the USA had the largest number of publications (1242), followed by England (274), Sweden (2267), and Canada (940) (Table 3). Among the top 10 countries/regions in PMS/PMDD research, there were two Asian countries, China and Japan. China was the only country from the developing world to be in the top 10 countries that contributed PMS/PMDD research, showing its vast progress in life science over the past decade.
Over 1700 institutions contributed to the publications on PMS/PMDD. Compared with countries, there was very little cooperation between the institutions (Figure 5). The lines between nodes represent the cooperative relationships among institutes. The length and thickness of the lines represent the degree of cooperation among countries. The top 10 institutions contributed to 570 articles, which accounted for 20.12% of the total number of publications. The UNIV PENN led the first research echelon, followed by the NIMH, the Univ Calif Los Angeles, UMEA UNIV, UNIV N CAROLINA, and HARVARD UNIV (Table 3).
Analysis of author and co-cited author
More than 6,640 authors contributed to the total number of publications. The cooperation between authors is presented in a network map (Figure 6). The size of the nodes represents the number of citations. For authors who had the most publications, Backstrom T ranked the first (105 publications), followed by Rubinow DR (59 publications), Freeman EW (49 publications), and Schmidt PJ (45 publications) (Figure 6 and Table 4).
Author citations are often used to estimate the scientific relevance of publications. The top ranked author by citation count is Halbreich U (1978), with a citation count of 890. The second one is Steiner M (1980), with a citation count of 699. The third is **the American Psychiatric Association (1986), with a citation count of 616. The 4th is Freeman EW (1988), with a citation count of 603. The 5th is Rubinow DR (1985), with a citation count of 492. The 6th is Yonkers KA (1996), with a citation count of 491. The 7th is Endicott J (1984), with a citation count of 467. The 8th is Schmidt PJ (1992), with a citation count of 463. The 9th is Rapkin AJ (1990), with a citation count of 428. The 10th is Backstrom T (1978), with a citation count of 393 (Table 4).
For the top 10 active authors with publications on PMS/PMDD, we can see that the authors that published the largest number of papers were not those whose citation count was the largest. In other words, high volume is not necessarily high yield. To do scientific research and publish papers, authors should pay more attention to quality rather than quantity (Table 4).
Analysis of co-cited references
We generated a cited reference co-citation map by selecting the top 10 articles per year and mapping them in 485 nodes and 241 links (Figure 8). An analysis in terms of co-citation counts (Tables 5 and Figure 8) revealed that the data on this topic over the past years were generally in the form of randomized trials, comparisons of diagnostic criteria (PSST), pathogenesis (GABA) and so on.
Analysis of Co-Occurring Keywords and Burst Terms
Over time, a knowledge map of keyword co-occurrence could reflect hot topics, and burst keywords (keywords that are cited frequently over a period of time) could indicate frontier topics. CiteSpace was used to construct a knowledge map of co-occurring keywords and identify the top 20 keywords in publications from 1945 to 2018 according to frequency, citation counts and centrality (Table 6). Generating a keyword co-occurrence map resulted in 150 nodes and 842 links (Figure 9). Among the listed keywords, "premenstrual syndrome, menstrual cycle, premenstrual dysphoric disorder, women, symptom, luteal phase, depression, dysphoric disorder, premenstrual symptom, prevalence, double blind, progesterone, mood, oral contraceptive and major depression" ranked ahead in both frequency and centrality, which suggested that they were the hotspots in the field. As we can see, "premenstrual syndrome and premenstrual dysphoric disorder" ranked in the top 3 keywords, which is reasonable because they are our search terms. Except for them, other key words actually reflect what we are focusing on.
According to the analysis of research hotspots, hot research topics revolve around the relationship of women's menstrual cycle with symptoms and progesterone. PMS/PMDD occurs only in women. According to the DSM-5, women with PMDD must have at least 5 predominantly affective symptoms with functional impairment, of which affective symptoms make up the largest proportion, such as mood swings, irritability, anger and depressed mood. Therefore, the relationship between symptoms of PMS/PMDD and the menstrual cycle has been a research hotspot. Many studies have explored this area[20-22]. Due to these findings, many scholars apply research in this area to disease subtype exploration. For instance, Chinese scholars proved that liver-qi invasion syndrome and liver-qi depression syndrome are the main subtypes of PMS/PMDD with epidemiological research[23, 24]. Regarding cyclic mood disorders, the follicular phase and luteal phase have been the research focus. According to existing research results, cognitive, sensory, and emotional changes are associated with the menstrual cycle. The reason for this association is the ovarian hormones, especially the hormones progesterone and estrogen, and their sophisticated fluctuations over the course of the human female menstrual cycle play a dominant role in the development of PMS/PMDD[25]. Progesterone and its metabolites (e.g., allopregnanolone) have been regarded as hot topics in scientific research on PMS/PMDD[26][27]. Recent studies now provide strong evidence that allopregnanolone could ameliorate the symptoms as a result of its ability to antagonize the allopregnanolone effect on the GABAA receptor[28].
So-called “burst words” represent words that are cited frequently over a period of time. CiteSpace was used to detect burst keywords, which are considered to be indicators of research frontier topics over time. In Figure 10, the time intervals are plotted on the green lines, while the periods of burst keywords are highlighted in red, indicating the beginning and end of the time interval of each burst. Among them, the keywords with citation bursts after 2008 were as follows: "prevalence" (2008-2018), "systematic review" (2009-2018), "impact" (2011-2018), "dysmenorrhea" (2011-2018), "confidence interval" (2011-2018), "menstrual cycle phase" (2012-2018), "risk factor" (2013-2018), "anxiety" (2013-2018), "postpartum depression" (2013-2018), "premenstrual phase" (2014-2018), "dysphoric disorder" (2014-2018), "control group" (2015-2018), "quality of life" (2016-2018), and "young women" (2016-2018). In the following, we list frontiers of PMS/PMDD research.
Prevalence and impact in young women:
Currently, it is estimated that 3-8% of women of reproductive age meet the strict criteria for PMDD[29]. The assessment of published reports demonstrated that the prevalence of clinically relevant PMS/PMDD is probably higher due to the strict diagnostic criteria. Although DSM-IV or DSM-5 are the main diagnostic criteria for PMDD, we found an interesting phenomenon in which researchers used a wide variety of diagnostic tools to determine the incidence of the disease in searching for cold-related literature. There are still many problems in the implementation of diagnostic criteria[30]. For one, patients need to report bothersome premenstrual symptoms, and clinicians should become more proficient in the diagnostic process to prevent the underdiagnosis of these disorders[31]. Estimating the prevalence of PMS/PMDD in greater scope and depth in different countries/regions is becoming increasingly important. In further studies, we found that young women are becoming the center of research, such as women in the Ukraine[32], in universities in Jordan [33], and in Japanese colleges [34]; Israeli students[35]; and other students all over the world. In recent years, a large number of related studies have endlessly focused on young women worldwide, such as young adult women[36], Israeli students[37], adolescent athletes[38], and Japanese collegiate athletes[39]. On the one hand, research output on the prevalence and impact of PMS/PMDD is also increasing around the world, such as consequences on female students’ behavior, cognitive abilities, mental health status, and academic performance[40] and the impact of symptoms on the quality of life[38]. However, researchers have not yet fully grasped the prevalence of the disease, lacking systematic reviews on prevalence.
Systematic Review evaluating risk factors
Systematic reviews typically involve a detailed and comprehensive plan and search strategy derived a priori, with the goal of synthesizing findings qualitatively or quantitatively[41]. Systematic reviews of randomized controlled trials are key to the practice of evidence-based medicine and to evaluate the effectiveness of drugs and methods for the treatment of uncomfortable symptoms. For example, cognitive-behavioral therapy[42], herbs[42], and vitamins and minerals are used in the treatment of premenstrual syndrome[43]. An review of fMRI to outline the neural basis of PMDD was performed in women across the reproductive lifespan and elucidated the role of neuroendocrine involvement in the development of depression in women[44]. A systematic review of acupuncture and acupressure showed improvements in both the physical and the psychological symptoms of PMS when compared to a sham control[45]. A systematic review of treatment pointed out the curative effect of herbal remedies for the treatment of PMS[46, 47].
Association of anxiety and depression with menstrual cycle phases
Menstrual-related mood disorders such as PMDD are mood disorders associated with the menstrual cycle. As major symptoms of many emotional disorders, anxiety and depression are the key directions of research. Moreover, some studies regarded PMS/PMDD as a risk factor for postpartum depression[48-50]. Symptoms of irritability, emotional hypersensitivity, increased anxiety and food cravings, sleep difficulties, and decreased concentration characterize PMDD as well as depression, particularly atypical depression. A lifetime history of depression ranges from approximately 20% to 76% in samples of women diagnosed with PMS or PMDD[51]. Recent studies have shown that menopause and menstrual cycle phases are times of intense hormonal fluctuation that can cause increased vulnerability to depression and anxiety[52].