The Relative Health Benefits of Different Sexual Activities

ABSTRACT

Introduction.  Although many studies examine purported risks associated with sexual activities, few examine potential physical and mental health benefits, and even fewer incorporate the scientifically essential differentiation of specific sexual behaviors.

Aims.  This review provides an overview of studies examining potential health benefits of various sexual activities, with a focus on the effects of different sexual activities.

Methods.  Review of peer-reviewed literature.

Main Outcome Measures.  Findings on the associations between distinct sexual activities and various indices of psychological and physical function.

Results.  A wide range of better psychological and physiological health indices are associated specifically with penile–vaginal intercourse. Other sexual activities have weaker, no, or (in the cases of masturbation and anal intercourse) inverse associations with health indices. Condom use appears to impair some benefits of penile–vaginal intercourse. Only a few of the research designs allow for causal inferences.

Conclusions.  The health benefits associated with specifically penile–vaginal intercourse should inform a new evidence-based approach to sexual medicine, sex education, and a broad range of medical and psychological consultations. Brody S. The relative health benefits of different sexual activities. J Sex Med 2010;7:1336–1361.

Introduction

Many studies have examined purported health risks associated with sexual activities, but few studies have examined the potential physical and mental health benefits of sexual activities. Even fewer studies have incorporated the scientifically essential differentiation of specific sexual behaviors. Sexual medicine should be concerned not only with the treatment of ill sexual health, but the specific evidence-based promotion of positive health.

An overview of the empirical evidence on psychological and physiological differences between sexual behaviors should inform the practice of sexual medicine, as well as other fields, including general medical practice, sex education, psychotherapy (both sexuality topics as well as sexuality as an aspect of character), and research in physiology and psychology.

Aim

This review aims to provide an overview of studies examining potential psychological and physiological health benefits or correlates of various sexual activities, with a primary focus on studies differentiating the effects of distinct sexual activities. An additional aim is a discussion of implications for education, research, and clinical assessment and practice.

Methods and Main Outcome Measures

Studies were identified from various sources, including those studies conducted in the laboratories of the author and colleagues, previous briefer overview papers incorporating some aspects of health benefits associated with sexual behaviors [1–3], and searches in PubMed and PsycInfo. The present review differs from those earlier reviews in its breadth and (with the exception of one [1]) greater attention to differences between sexual behaviors. There is also more discussion of methodological issues, as well as a discussion of implications for education, research, and clinical assessment and treatment. The presentation is organized largely by the nature of the health topics, which are found under the major headings of Psychological and Physiological factors (some topics, such as pain, could have arguably been placed under either major heading). The topics range from “soft” variables such as relationship quality to “hard” variables such as biochemical measures. The nature of the research designs varies greatly. Although experimental designs generally are viewed as offering firmer evidence than correlational or risk-factor epidemiological designs, the reader is advised to look for convergent evidence using differing samples and methods, optimally with examination of some potential confounding variables.

A few studies noted in this review adjust for the tendency of some people to underreport behaviors that they consider to be socially undesirable. Although it might be conjectured that elevated social desirability response scores might measure a propensity to actually behave in a socially desirable manner, there is some evidence to the contrary: social desirability scores were higher than normative values in large groups of men and women convicted of intimate partner abuse [4]. Social desirability response bias scores have been found to mask the ability of greater life stress and poorer psychological coping skills to predict athletic injuries [5]. Social desirability response bias scores predict discrepancies between (indirectly) measured and self-reported caloric consumption [6–8]. However, social desirability responding varies between individuals and situations, and as such, is not always a confounding factor. Interestingly, social desirability responding might itself be associated with predictors of poorer physical health outcomes [9,10].

Most cross-sectional studies that focus on infirm populations have been excluded, because of the risk of reverse causality (inability to engage in some sexual behavior because of the infirmity). In contrast, the review includes studies that find sexual behavior differences in healthy adults with regard to subtle but important longitudinal predictors of future morbidity or mortality (such as heart rate variability, and blood pressure stress reactivity). However, the issue of causal direction (including bi-directional causality, as in vicious or virtuous circles) needs to be pondered in any correlational or risk-factor epidemiological research design. Retrospective case-control studies might be subject to biased recall and/or insufficient matching of cases and controls. Associations can also be a result of a shared unmeasured third factor (such as genetic influence) that affects both the nominal outcome variable and the nominal predictor variable. These issues apply to risk-factor epidemiological studies in general, not only to studies with potentially controversial findings.

Differences between health aspects of specific sexual behaviors (e.g., penile–vaginal intercourse [PVI], masturbation, sex with a partner other than PVI) are highlighted. Some of the studies examine whether one has engaged at all in a sexual behavior in a given time period, others examine frequency, others examine combinations of sexual repertoires, others examine manner of orgasm elicitation, and others examine gross modification of the nominal sexual behavior (e.g., condom use or clitoral masturbation during PVI).

The discussion will address the compatibility of the observed results with both evolutionary and early psychoanalytic theories. These two approaches to understanding human behavior focus respectively on which characteristics have been most adaptive in the course of human evolution (in the sense of increasing the likelihood of transmitting one's genes), and the mental operations (many out of awareness) that motivate behavior and are rooted in problems of childhood psychosexual development (with implications for chronological adults). The discussion also describes some physiological and other possible bases for the observed effects.

Results

Psychological Factors

Satisfaction with One's Mental Health

In a large representative sample of the Swedish population, PVI frequency was a significant predictor of both men's and women's greater satisfaction with their mental health [11]. In contrast, masturbation was inversely associated with mental health satisfaction in the multivariate analyses that controlled for other sexual behavior frequencies, and partnered sexual behaviors other than PVI were uncorrelated with mental health satisfaction [11]. The same large Swedish survey also revealed that women who had experienced vaginal orgasm (defined quite conservatively as having “had an orgasm solely through the movement of the penis in the vagina”) were more satisfied with their mental health than the minority of women who had only experienced orgasms through direct clitoral manipulation [12]. The Swedish survey used in those studies was reported to not be affected by social desirability response bias [13,14].

Intimate Relationship Quality

In a small study of Portuguese women, frequency of PVI (FSI) correlated positively with Perceived Relationship Quality Components (PRQC) Inventory dimensions of satisfaction, intimacy, trust, passion, and love [15]. In contrast, frequency of partnered sexual behaviors other than PVI was uncorrelated with the PRQC dimensions. Masturbation frequency was inversely associated with love. PVI orgasmic frequency correlated positively with the PRQC dimensions of satisfaction, intimacy, passion, and love. PVI orgasmic consistency (proportion of PVI occasions resulting in PVI orgasm) was inversely associated with masturbation frequency. When PVI frequency was controlled in a partial correlation procedure, noncoital sex frequency was associated with less global relationship satisfaction, and noncoital partnered orgasm frequency was associated with less love. Social desirability scores did not confound the associations [15]. These results are fairly consistent with an American study which found that 100% of maritally and sexually satisfied wives—but only 68% of the maritally satisfied yet sexually dissatisfied wives—had PVI among their sexual activities in the past week, that masturbating the male to orgasm was reported by 4% of the maritally and sexually satisfied but 30% of the maritally and sexually dissatisfied women, and that cunnilingus frequency was unrelated to satisfaction [16]. Although not emphasized in Kinsey's writings, Kinsey researchers found that “marital happiness” was associated with female coital orgasm [17]. In a large representative sample of the Swedish population, independent multivariate predictors of men's relationship satisfaction were greater frequency of PVI, but lesser frequencies of masturbation, anal sex, and oral sex (for women, the independent multivariate predictors were simply greater frequency of PVI and lesser frequency of masturbation) [11].

Alexithymia

Alexithymia is a relative inability to perceive, identify, and express emotions. It is a personality trait associated with some forms of psychopathology, associated with the use of immature psychological defense mechanisms and also associated with the use of distraction as a coping mechanism [18]. Studies of Swiss and American patients found that alexithymia was associated with hypoactive sexual desire [19], sexually dysfunctions, and paraphilias [20].

FSI (but not frequency of either masturbation or of partnered sexual activity excluding PVI), as measured by both sexual behavior diaries and recall, was associated with less alexithymia (hence, more emotional integration) as measured by the Toronto Alexithymia Scale (TAS-20) in a sample of healthy German women [21]. High social desirability scorers were excluded from the analysis.

In addition to the use of the validated psychometric measure of alexithymia, the association between habitual sexual behaviors and the laboratory measured integration of vaginal responses into psychological arousal can inform the understanding of emotional integration. Research has revealed that the awareness and integration of vaginal sensations into women's subjective sense of sexual arousal varies as a function of their habitual orgasm sources.

Several studies have concurrently examined women's subjective and vaginal response to erotica (videotape and/or fantasy), and in contrast to studies with men, most of the studies showed that overall, there was poor concordance between women's vaginal and subjective sexual arousal responses. Based on the inference that women who have PVI orgasms were more likely than coitally anorgasmic women to successfully integrate their physical and emotional feelings, it was hypothesized [22] that greater orgasmic consistency during PVI (but not during other sexual behaviors) would be associated with better concordance of vaginal and subjective sexual arousal.

Healthy menopausal Dutch women completed a questionnaire on their PVI, masturbation, and noncoital partner sexual frequencies for a 1-month period, and noted for each occasion whether orgasm occurred; orgasm consistency was the percentage of each sexual event type resulting in orgasm [22]. In the laboratory, they were exposed to erotic videotapes, nonerotic (control) videotapes, and also asked to engage in epochs of sexual fantasy. They rated their sexual arousal, and their vaginal response was measured with a vaginal photoplethysmograph that assesses vaginal vasocongestion (the device allowed measurement of vaginal pulse amplitude). The correlation (z-transformed) of subjective and vaginal response was the index of concordance. As hypothesized, concordance was significantly associated with PVI orgasm consistency, but not consistency of orgasm during other sexual activities: women who regularly had PVI orgasms had excellent concordance of vaginal and subjective arousal, but other women (even those who orgasmed reliably through means other than PVI) had a functional disconnection between their vaginal arousal and their mental experience. The results were not confounded by social desirability responding.

Also of note were the findings that (i) orgasm consistency rates were similar for PVI and for noncoital sexual activities, and (ii) PVI and masturbation orgasm consistency were uncorrelated. This latter finding implies that most of the coital orgasms were most likely not masturbatory clitoral orgasms, but real vaginal orgasms (i.e., female orgasm induced by penile–vaginal stimulation per se). The latter finding also has implications for sex therapy, as masturbation and intercourse orgasms are substantially different.

The same pattern and magnitude of vaginal-subjective arousal concordance results were found in a replication study [23] involving young Dutch women (all of whom had current partners; in the first study a few of the women did not have current partners).

It is specifically PVI orgasm consistency that is related to integration of vaginal response into the appraisal of arousal. An analogy might thus be made between these findings and the aforementioned ones on alexithymia: in both cases an index of specifically and exclusively PVI reward (frequency [21] or orgasm consistency) was associated with an index of greater awareness of feeling (vaginal sensation or differentiated emotions).

Immature Psychological Defense Mechanisms

Psychological defenses are processes, generally operating outside awareness, that reduce distress caused by emotional conflict. Immature (maladaptive) defense mechanisms involve a distortion of reality and/or impairment of awareness, and they are associated with a variety of indices of poorer mental health and relatedness, including psychological immaturity and lesser ability to relate intimately with the opposite sex [24,25]. Immature defense mechanisms are associated with a variety of psychiatric disorders [26–29]. According to early psychoanalytic theories, psychological immaturity (psychosexual immaturity, with its concomitant greater use of immature defense mechanisms) could lead to inhibition of frequency and appreciation (including vaginal orgasm) of PVI in favor of other or no sexual behaviors, with noxious consequences for mental health and intimate relationships.

In a sample of healthy Portuguese women, vaginal orgasm (triggered solely by PVI) was associated with less use of immature defenses [24]. Defenses were measured with the Defense Style Questionnaire, a well-validated (including association with various psychopathologies) measure of immature, neurotic, and mature defenses [26,30]. Vaginal orgasm was associated with less overall use of immature defenses, as well as with less use of the specific component immature defenses: somatization, dissociation, displacement, autistic fantasy, devaluation, and isolation of affect. Orgasm from clitoral stimulation or combined clitoral–intercourse stimulation was not associated with less use of immature defenses, and was associated with more use of some immature defenses (e.g., orgasm from noncoital partner activity in the past month was associated with the defense of dissociation). In one multivariate analysis, both (i) any masturbation orgasm in the past month and (ii) less vaginal orgasm consistency, made independent contributions to the statistical prediction of immature defenses. In another regression analysis, (i) any use of extrinsic clitoral stimulation for intercourse orgasm and (ii) lack of any vaginal orgasm, made independent contributions to the statistical prediction of immature defenses. Vaginally anorgasmic women had immature defenses scores comparable to those of established (depression, social anxiety disorder, panic disorder, and obsessive–compulsive disorder) outpatient psychiatric groups. Results were not confounded by social desirability responding [24].

A study of predominantly Scottish women who completed an anonymous Internet-based survey [31] provided a cross-cultural replication of the results obtained in the Portuguese sample. Greater use of immature psychological defense mechanisms was associated with lesser vaginal orgasm consistency, with any orgasm from clitoral masturbation during PVI, and with greater frequency of masturbation orgasm. Immature psychological defense mechanisms were also associated with greater frequency of masturbation during PVI, and with frequency and orgasm frequency of both anal sex and vibrator use. Immature psychological defense mechanisms were also associated with greater quantity of alcohol consumed before sex [31].

Condoms impair many aspects of PVI, including intimacy and sensation [32]. Freud opined that condom use during PVI, like sexual activities other than PVI, led to a detrimental effect on orgasm that fueled the neuroses. More psychologically immature people might prefer condoms for PVI as a means of reducing intimacy and/or such reduction might hinder psychological growth. Indeed, research indicates that condom users do have poorer relationship quality with their partners [25,33]. In a study of healthy Portuguese adults, frequency of PVI with condoms correlated directly with use of immature defenses. In contrast, frequency of PVI without condoms correlated inversely with use of both immature and neurotic defenses. Results were not confounded by relationship status, age, cohabitation, or social desirability responding. Regression analyses revealed that immature defenses were independently predicted by condom use for PVI and by masturbation orgasms (for both sexes). For women, additional predictors were orgasm from clitoral masturbation during PVI, and lack of vaginal orgasm. The results are consistent with condom use during PVI being associated with psychological immaturity and predisposition to poorer mental health [25].

Depression

Higher masturbation frequency (and even the desire for more masturbation) is associated with depression [34–36], and masturbation is associated with less happiness [37]. The association of masturbation with depression is unlikely to be a result of simply a lack of PVI, because more masturbation and less PVI make independent contributions to less satisfaction with relationships, sex life, life in general, and one's mental health (the multivariate analyses also examined some partnered sexual activities other than PVI, and revealed that anal and oral sex frequency also have independent inverse associations with some of the satisfaction indices) [11].

It is likely that only unfettered, real PVI has important mood-enhancing benefits. A study of young women in the United States found that not only did Beck Depression Inventory scores worsen with increasing time since last PVI (i.e., lower FSI is associated with more depression), but the use of condoms obliterated the apparent antidepressant effects of PVI [38]. Depressive symptoms and suicide attempts among women who used condoms were proportional to the consistency of condom use: more condom use means more depression and more suicide attempts. The depression and suicide association with condom use was not because of confounding by relationship duration, and there were no differences in depression between those in and not currently in a relationship. The investigators suggested that their results (the results were reportedly replicated, but the details of the replication have not yet been published in detail [39]) might be a result of intravaginal absorption of seminal prostaglandins (as well as possibly seminal testosterone, luteinizing hormone [LH], and oxytocin) improving the mood of women, but the researchers did not measure the relevant chemicals in the individual research participants. Even among women who rarely or never used condoms, depressive symptoms were associated with urinating after intercourse [39], which the investigators noted would decrease the quantity of seminal components that could be absorbed. Compared with women who did not use condoms at all, consistent condom users were both more depressed at baseline and also evidenced a worsening of their mood during a longitudinal study in the United States [33]. Although there might be a direct chemical antidepressant effect of semen absorbed from the vagina, the large difference in mood and suicidality might also be a result of intercourse with condoms not really being intercourse, but something akin to mutual masturbation with the same latex device.

One group that has both elevated rates of masturbation [40] and of partnered sexual activities other than PVI is homosexuals. Large representative surveys have found much higher rates of suicidal ideation, mood disorders, substance use disorders, and other psychiatric disorders in homosexual men and homosexual women than in their heterosexual counterparts (including in studies conducted in what is probably one of the most pro-homosexual countries: The Netherlands) [41–43]. Although it was noted that among homosexual men, perceived discrimination was associated with suicidality [43], it was not made clear that the defensive process of attributing one's bad feelings to other people is itself part of the process of depression for some people [44]. In a large UK-based survey, 61.7% of homosexual men who had sex with a man in the past year reported receiving ejaculate in their mouth (although 97.2% reported performing fellatio at all) and 42.2% reported receptive anal intercourse without a condom in the past year [45] (unfortunately there was not more precise quantification of frequency of the behaviors, except that in the updated version of the survey, 24% reported receptive anal intercourse without a condom in the past month, and an additional 18.5% reported it over a month but less than a year ago [46]). Thus, most homosexual men (who were active in the past year) appear to have the opportunity for absorption of seminal components at some location within the alimentary canal. The combination of high rates of depression despite the opportunity for absorption of seminal components suggests explanations including some combination of: site-specific or sex-specific effects of seminal component absorption (i.e., vaginal absorption has important antidepressant effects not afforded by alimentary absorption), variations in frequency of seminal exposure, or overwhelming genetic or psychosexual developmental effects linking homosexuality and depression. A recent large twin study revealed that not only do nonheterosexual men and women have significantly higher neuroticism and psychoticism scores than heterosexuals (implying elevated psychiatric risk), but also that there is a significant genetic correlation between nonheterosexuality and both neuroticism and psychoticism, but no significant environmental correlation with nonheterosexuality. This implies that any common cause of both nonheterosexuality and psychiatric risk is likely to be genetic rather than environmental [47].

Other Psychiatric Disorders In a study [48] comparing Scottish schizophrenics with controls of the same age, sex, and postal code, it was found that the schizophrenic men were several times more likely to report a zero FSI than were controls, and less likely to report intercourse of at least weekly. However, they were not more likely to report abstaining from masturbation or masturbating at least once weekly. Similarly, schizophrenic Scotswomen in the study were more likely to report a zero FSI than controls, but they did not differ in their masturbation frequency from controls.

Although some antipsychotic medications might cause sexual dysfunction and low desire (for some, at least in part through prolactinergic mechanisms) [49], there is also evidence that untreated schizophrenics have low desire for sex with a partner, and other studies also indicate that schizophrenics have low rates of partnered sexual activity but elevated rates of masturbation [49]. Besides the issue of social skills deficits, the low PVI frequencies among schizophrenics might be in part the result of wanting to avoid intimacy. Apparently the tendency toward anhedonia (inability to experience pleasure) among schizophrenics does not limit masturbation, only PVI pleasure. The presence of another person during sexual activity, particularly for PVI, the most emotionally intimate sexual activity, might be aversive to many schizophrenics. Failure to develop a stable integrated self might also impair appreciation of PVI, particularly PVI orgasm. These aspects of the schizophrenic situation might be viewed on a continuum with the rest of the population, such that there are many nonpsychotics who are would be overburdened by the emotional intimacy and overwhelming pleasure that PVI can provide.

Czech female schizophrenics, manic-depressives, neurotics, anorexics, and a control group of gynecological spa patients were interviewed regarding their sexual histories [50]. The manic-depressive (bipolar) patients did not differ from the control group in their prevalence of PVI orgasmic response, but the schizophrenics, neurotics, and anorexics all had lower rates of coital orgasm. Thus, it was not only the schizophrenics, but women with some quite different forms of psychological problems that were impaired in their ability to orgasm from intercourse.

Similarly, women with neurotic disorders were less likely to have PVI orgasms than were a group of women without neurotic disorders. However, they were no less likely to have orgasms from direct clitoral stimulation [51].

A similar pattern was found in a Czechoslovak study comparing prostitutes with nonprostitutes [52]. The prostitutes were less likely than the control group to orgasm during PVI. Prostitutes (especially street prostitutes) have a high prevalence of various forms of psychopathology, including antisocial personality disorder, borderline personality disorder, dissociative disorder, depression, schizophrenia, and other psychiatric and personality disorders [53]. This psychopathology, which includes psychological processes dissociating the experience of PVI, might contribute to the high mortality rate of prostitutes [53,54].

Successful nonhormonal treatment of erectile dysfunction led to an increase in FSI, a concomitant decrease in masturbation frequency, and a decrease in psychiatric symptoms [55].

Physiological Factors

Analgesia and Pain

Prostatodynia is characterized by urinary symptoms and pelvic pain suggestive of prostatitis but with a nonpathological prostate examination and without signs of inflammation or infection in prostatic secretions. In a report on prostatodynia [56] in United Nations peacekeeping forces, the occurrence of the disorder was associated with not having PVI, and it resolved with recommencement of PVI. However, masturbation led to either no improvement or to an exacerbation of pain symptoms.

Vaginal stimulation has been shown to have substantial analgesic properties, far greater than clitoral stimulation [57]. The effects are not attributable to distraction [58], and appear to not be a result of an opiate-type process [59]. One risk for women with some sexual pain disorders is the learned avoidance of the potentially most analgesic and psychologically fulfilling sexual behavior.

Vaginal and Pelvic Muscle Function

A review of the literature on female genito–genital reflexes concluded that vaginal intercourse helps to maintain vaginal and pelvic function, including through penile thrusting triggering reflex muscular contractions that maintain and improve vaginal function [60]. There were also indications that the presence of seminal component prostaglandin PGE1 in the vagina after ejaculation might maintain vaginal oxygenation and blood flow. Improving blood flow could be expected to support sexual response and vaginal health (and perhaps general health). Using condoms deprives women of many benefits, including those blood flow and oxygenation benefits.

Functional Musculoskeletal Disturbance

Some theories of personality and psychotherapy have proposed a link between chronic muscle blocks and disturbances of both character and sexual function [61]. Regardless of whether the sites of chronic muscle blocks (or chronic muscle flaccidity) have metaphorical meaning, they might be both an indicator and mechanism for impaired function, including sexual function. A study examined the association of general everyday body movement with history of vaginal orgasm by asking healthy young Belgian women with known histories of either vaginal orgasm or vaginal anorgasmia (50% from each group) to be videotaped walking on the street; their vaginal orgasmic status was judged by trained (in the Functional-Sexological school) sexologists blind to their history [61]. History of vaginal orgasm was diagnosable at far better than chance level (81.25% correct). The gait of the vaginally orgasmic women was characterized by being physiologically normal, and manifested fluidity, energy, sensuality, freedom, and absence of both flaccid and locked muscles (greater pelvic and vertebral rotation were characteristic of the vaginally orgasmic women). Clitoral orgasm history was unrelated to both vaginal orgasm history and vaginal orgasm rating. The report also noted previous studies that differentiated homosexual and heterosexual men and women on the basis of other aspects of gait, and discussed functional musculoskeletal issues, the effect of the musculature on sexual function, and implications for sexual therapy [61].

Metabolism and Nutrition A study of healthy German adults revealed that a (physician-measured) slimmer waist (for men and the sexes combined) and slimmer hips (for men and women) were each associated with greater PVI frequency [62]. In contrast, slimmer waist and hips were associated with lesser masturbation frequency (men and the sexes combined), and noncoital partnered sexual activity had a less consistent association with slimness. Waist and hip circumference were associated inversely with PVI importance for men. Cohabitation status was an independent predictor of PVI frequency, and did not confound the association of slimness with sexual behavior. Of note, the effects were obtained despite exclusion of obese and medically unfit subjects from the study. Persons with high social desirability scores were also excluded from the analyses.

In addition to issues of attractiveness (likely rooted in evolutionary processes that favor healthier partners), higher body fat levels are associated with lower testosterone levels and with less brain dopamine activity, and overeating tends to increase brain serotonergic tone. These factors might impair sexual desire and/or function, especially the most complex and evolutionarily relevant sexual behavior: PVI [62]. Similar differential sexual behavior associations with slimness have been observed in other species: when presented with slim females, both slim and obese male rats hold the females and lick the vaginal region, but obese males have a much lower PVI frequency than the slimmer males [63]. Although in some human societies, there is a tendency for men to find heavier women more attractive, these societies tend to be characterized by food scarcity (or limited food storage), making the issue of energy storage within the potential fetus bearer more evolutionarily salient than the issue of the various morbidities associated with overweight (factors that become more salient when food scarcity is not a major issue) [64].

In societies in which food scarcity is not a problem, people who enjoy PVI might make efforts to remain sexually desirable to partners by staying slim. Studies have indicated that adults who report having been physically or verbally abused in childhood or adolescence are more likely than others to be overweight or obese [65]. More importantly, a review of longitudinal studies reported that hostility, anger, and depression longitudinally predict the development of higher levels of adiposity (and other aspects of the metabolic syndrome) [66]. It would not be surprising if a history of dissatisfaction with intimate relationships leads to self-destructive and intimacy avoiding behaviors (including specifically minimizing PVI frequency).

Ascorbic acid (vitamin C) has many functions (including reduction of approach anxiety, modulation of brain dopaminergic and noradrenergic activity, cardiovascular support, oxytocin secretion, and reduction of stress [67,68]), some of which might support sexual behavior and some of which might only be manifest at high doses. A double-blind randomized controlled trial of high-dose (14 days of 3,000 mg/day sustained release) ascorbic acid in healthy young German adults led to the finding that ascorbic acid caused an increase in PVI frequency, but not in frequency of masturbation or of partnered sexual behaviors other than PVI [69]. The ascorbic acid also improved mood. The results were not confounded by social desirability responding. Exploratory analyses revealed that the effect was largely a result of the response of women in the study. However, the seeming sex difference might be attributed to young women perhaps being more likely than young men to fulfill rapidly any increased desire (caused by more optimal nutrition) within the 14-day recording period.

Cardiovascular Health

Resting heart rate variability (HRV, an index of autonomic cardiac regulation) reflects heart rate fluctuations in response to subtle homeostatic demands, including the differing autonomic effects of inspiration and exhalation. Resting HRV is driven largely by the parasympathetic nervous system (particularly the vagus nerve), which also has a role in sexual arousal. HRV is longitudinally predictive of lower mortality rates among both normals and persons with a history of heart disease [70,71]. Psychophysiological studies found that greater HRV is associated with indices of better mood, attention, self-regulation, and responsiveness to emotional experience. HRV is also associated with fewer antisocial personality features, less sexual dysfunction, and perhaps with pair-bonding processes [70,71]. In healthy German adults, greater HRV was associated with greater PVI frequency, but not frequency of either masturbation or of noncoital partnered sex [71]. HRV was also associated with greater subjective importance of PVI. These results were obtained after exclusion of persons with high social desirability scores. The results were replicated in a larger healthy German sample [70] and also found to be unconfounded by the various candidate variables that were examined. There were no sex differences in these statistical relationships. Thus, specifically PVI but not other sexual behavior was associated with an important measure of better homeostasis, better parasympathetic tone, lower mortality risk, and better psychological function (including better relatedness). Of course, given the correlational design, it is unclear to what extent better HRV leads to more PVI and appreciation thereof, to what extent more PVI leads to better HRV, and to what extent some other factors (such as genetics) influence both HRV and PVI.

In the first of the two HRV studies [71], but not the second [70], greater PVI (but not other sexual behaviors) was also associated with lower resting diastolic blood pressure. In an Italian study, newly diagnosed, never treated hypertensive married men aged 40–49 had an FSI 25% lower than a control group of men who had blood pressure in the normotensive range [72].

Blood pressure reactivity to acute stress is a risk factor for the development of hypertension and left ventricular hypertrophy, as well as for myocardial infarction or death in susceptible persons. It is mediated by increased sympathetic and decreased parasympathetic nervous system activity (the latter being largely a withdrawal of vagus activity). In a study of healthy German adults, blood pressure was recorded at baseline, after being instructed to prepare a speech, after giving the speech to an unsupportive audience for 5 minutes (plus 5 minutes of time-pressured verbal arithmetic), and after a 10-minute recovery period, as per the protocol of the standardized Trier Social Stressor Test [73]. The blood pressure responses were analyzed as a categorical function of whether participant diaries indicated that they engaged at all (in the past 14 days) in PVI, masturbation, or partnered sexual activity in the absence of PVI the same day [73]. Persons with high social desirability scores were excluded from the analyses. Persons who engaged in PVI but no other sexual behaviors (except partnered sexual activity the same day as PVI) during the fortnight had not only statistically significantly less, but markedly less, blood pressure increase to the stressor than did persons engaging in other or no sexual activities. They also had faster recovery from the stressor (which some researchers suggest might be even more important for health than more moderate stress reactivity). Persons who engaged in PVI but also engaged in either masturbation or exclusively noncoital partnered sexual activity on other days had results intermediate between the PVI-only group and the noncoital partnered sexual activity only or masturbation only groups (who did little or no better than the no-sex group) [73]. Overall, masturbation frequency was correlated with higher (worse) blood pressure reactivity.

Results were not confounded by any of the many demographic, physiological, psychological, or behavioral (including exercise minutes per week) candidate variables that were evaluated. The blood pressure reactivity difference between different sexual behavior groups was at least numerically greater than differences associated with the many psychological, physiological, social, and behavioral variables that have been examined in the sizeable psychophysiological literature on blood pressure stress reactivity in normals[73].

The pattern of results suggest not only that it is specifically PVI (rather than other sexual behaviors) that is associated with optimal cardiovascular “protection” from stress, but also that the benefits are not simply due to having a partner. The study indicated that engaging in masturbation or even partnered sexual behavior that excludes PVI on some days detracts from the benefits of PVI. There are several possible reasons for this effect, including: (i) partial (rather than complete) avoidance of PVI in favor of other sexual activities, (ii) approaching PVI with a psychological set akin to the other sexual activities, and (iii) the beneficial effects of PVI being partially offset by untoward effects of the other sexual activities.

A within-subjects laboratory study [74] compared the exercise value (oxygen uptake, and blood pressure and heart rate stimulation) of men engaged in: masturbation, masturbation by a partner, and two intercourse positions (man on top and woman on top). PVI produced greater duration of heart rate elevation and substantially greater oxygen uptake at orgasm than the other activities [74]. Although the exercise value was greatest for male on top (probably at least in part because of greater motor activity), the woman on top PVI was still superior to either of the manual forms of penile stimulation.

Pre-Eclampsia

Among the several studies reporting on the association between pre-eclampsia and less (unfettered) PVI is the report concluding that “birth control methods that prevent sperm exposure may play a role in the etiology of pre-eclampsia” (p. 3143) [75] and those reporting an odds ratio of 17.1 for pre-eclampsia associated with the combination of <4 vs. >12 months of “cohabitation” in combination with barrier contraceptive use [76]. Fellatio is also associated with decreased risk of pre-eclampsia: 44% of pre-eclamptic women, but 82% of controls reported fellatio before the index pregnancy [77]; however, that study did not assess frequency of PVI.

Hormonal and Neurohormonal Effects

In addition to any effects of testosterone levels supporting PVI frequency, there is evidence of the causal arrow pointing in the other direction [78]. Successful restoration of erectile function, through any of a variety of forms of nonhormonal treatment (psychological, surgical, vacuum device, yohimbine, prostaglandin E1) led to a return of testosterone levels to the typical range for the population. Men who responded to treatment with an FSI of at least 8 per month showed greater testosterone increases than men who responded with an FSI of 1–7, who in turn showed greater testosterone levels than nonresponders [78]. As the authors noted: “Since the pre-therapy low testosterone levels were independent of the aetiology of impotence, we hypothesize that this hormonal pattern is related to the loss of sexual activity, as demonstrated by its normalization with the resumption of coital activity after different therapies. The corollary is that sexual activity may feed itself throughout the increase in testosterone levels” (p. 385). The same research group found that the FSI response to sildenafil or tadalafil treatment also led to a corresponding decrease in serum LH [79]. These latter results need to be considered in light of recent experimental evidence that chronic administration of sildenafil itself stimulates testosterone secretion in mice [80].

In a randomized trial of coitus (vs. no coitus) for the treatment of Nigerian women's menopausal hot flashes, PVI was found to decrease hot flash symptoms, an effect perhaps mediated by the observed changes in LH and follicle-stimulating hormone (FSH) [81].

German women reporting earlier age at first intercourse (in all cases, postpubertal) had less intense cortisol increases in response to a standardized laboratory stressor (a nonsignificant trend in the same direction was observed for males), and faster recovery from the stressor [82]. In contrast, age at first intercourse was unrelated to the cortisol response to low-dose adrenocorticotropic hormone (ACTH) stimulation. Although shared genetic factors might influence both age at first intercourse and cortisol response to stress (with its many important untoward psychological and physiological sequelae), cortisol might impair dopaminergic function, thereby impairing initiation of and response to sexual activity [82].

PVI leads to increases in dopamine levels in the nucleus accumbens (a region important in at least attention to reward-related stimuli) of female hamsters. The brain stimulation effect was absent in females who were mounted as often by males but did not manage to have intercourse because of a hamster chastity belt: tape placed over their vaginal openings [83]. The chastity belt still allowed the females whatever benefits are afforded by being held and having stimulation of their external genital region by the males mounting them, but that was insufficient to trigger the brain stimulation effect provided by PVI. Thus, being vaginally stimulated by a penis provides substantially greater stimulation of the nucleus accumbens than that provided by partner masturbation of the external genital region.

The postorgasmic prolactin surge is associated with reduction of sexual drive and with some aspects of sexual satiety (whether directly through inhibitory central dopaminergic and peripheral processes, or as a process secondary to dopaminergic effects) [84,85]. Dopaminergic signals from the hypothalamus are a primary determinant of prolactin release, and dopaminergic neurons (including mesolimbic dopaminergic neurons) in turn can be modulated by prolactin; a lengthier discussion of these issues and a broader review of studies of prolactin control of sexual drive is available elsewhere [84,85]. In an experimental examination of different sexual activities in the laboratory by healthy adults, PVI caused both sexes to manifest a postorgasmic prolactin increase 400% greater than that following masturbation climax (adjusted for prolactin changes in a nonsexual control condition) [86]. The results indicate that PVI is not only more physiologically satisfying than masturbation, but the greater homeostatic dopamine modulating effects might be among the mechanisms involved in the psychological and physiological benefits associated with PVI rather than with other sexual activities [86], and might also underlie differences between the sexual satisfaction associated with various sexual activities [11,12,87].

Biochemistry of Ejaculate

In a sample of men with erectile dysfunction, greater frequency of masturbation was associated with greater prostate specific antigen levels (masturbation was also associated with prostate abnormalities, including a swollen or tender prostate) [88].

A comparison of semen ejaculated by the same men from masturbation and from PVI revealed that the volume of seminal plasma, sperm count, sperm motility, and percentage of morphologically healthy sperm were all greater in the PVI samples than in the masturbation samples [89]. In addition, markers of the secretory function of the prostate were significantly better for the PVI samples, which led the authors to infer that PVI was associated with a better prostatic secretory function than masturbation. Similarly, other researchers [90] also used a within-subjects design and found that compared with masturbatory samples of ejaculate, men's PVI samples had a larger semen volume and increased concentrations and total amounts of prostaglandin E and polyamines (putrescine, spermidine, and spermine). Thus, PVI involved better prostate function, larger semen volume, better quality sperm, as well as elimination of more waste products.

In another study [91] that found that PVI yielded far better sperm volume and sperm quality than did masturbation (an effect which was most dramatic in men with seminal deficiencies), the authors noted that although the greater sperm volume in the intercourse samples could be explained by greater sexual stimulation leading to “greater loading of the vas deferens prior to ejaculation” (p 192), they could not explain the basis for the difference in sperm morphology (quality), which is “determined prior to spermatozoa reaching the tail of the epididymis.” Thus, the organism appears to anticipate that there will be PVI rather than merely masturbation, and creates a higher quality seminal product, among other physical benefits.

The experimental finding of better prostate function being associated with PVI (rather than masturbation) is consistent with epidemiological studies linking PVI but not masturbation with fewer disorders of the prostate. It is also consistent with psychoanalytic and similar perspectives that intercourse is more “disburdening” than other sexual behaviors. Although this latter concept was largely meant to refer to the relief of tension, it appears to apply to other physiological domains as well. Rather than being just abstractions, some of the differences between PVI and other sexual behaviors can be measured in a test tube.

Prostate Cancer

Several studies have reported an association between recalled frequency of ejaculation and lower prostate cancer risk [92,93]. Unfortunately, some of these studies vitiated the opportunity to differentiate between protective and nonprotective ejaculation conditions [94]. A British study found no association of FSI with subsequent prostate cancer risk, but did find an increased prostate cancer risk associated with higher masturbation frequency in the 20s, 30s, and 40s, but the opposite effect in the 50s. The authors noted that the latter effect could be a result of, at least in part, reverse causation [95].

A review of the literature on sexual risk factors for prostate cancer found that cases had a lower FSI from age 50 than did age-matched healthy controls [96]. The cases also had a higher lifetime prevalence of sexually transmitted disease, more use of prostitutes, and a higher masturbation frequency [96]. Similarly, another study found that compared with controls, cases had a lower lifetime FSI (men who had intercourse more than 3,000 times in their lives had half the risk of those that did not), more homosexual partners (thus, not PVI), and more use of prostitutes [97]. The examination of prostitute contact as a separate category is important and commendable. It is not only a proxy for possible exposure to sexually transmitted disease (the usual interpretation), but a potential indicator of dissociated sexuality by both participants.

As noted in the section on biochemistry of ejaculate, there are important differences between sexual behaviors in markers of prostate function and quantity of waste products, and these differences could conceivably impact prostate cancer risk.

Breast Cancer

In a retrospective case-control study, both low PVI frequency and greater condom (or withdrawal) use were identified as risk factors for the development of breast cancer [98]. Women with infrequent or no PVI had thrice the breast cancer risk of the controls (women who had more frequent PVI) in the study. In addition, those women who did have PVI but used a contraceptive method that decreased pleasure by decreasing the contact of the vagina with the penis, and decreased the woman's vaginal contact with semen (women who used condoms or coitus interruptus), were at greater risk for developing breast cancer than women who used oral contraception or an intrauterine device [98]. The interactive effect (PVI frequency and not disrupting PVI by choosing condoms or withdrawal) was noteworthy: those women who had PVI (without the disruptive effects of condoms or withdrawal before ejaculation) for at least 20 years had one-tenth the breast cancer risk of women who never had intercourse. Once again, condom use (or withdrawal) is associated with the promotion of a life-threatening condition.

In addition to that French study, an earlier study from the United States found a similar association [99,100]. Women whose men used condoms or withdrawal (which generally involves the male switching from intercourse to masturbation for ejaculation), as well as women not engaging in intercourse at all, had a breast cancer rate five times higher than users of contraceptives that do not reduce vaginal exposure to semen.

Greater lifetime number of sexual partners was associated with decreased breast cancer risk [101], and nuns were found to have very high rates of breast cancer [102]. As noted in the Introduction, a variety of methodological factors should be considered in evaluating risk-factor epidemiological studies.

Immune Function and HIV

Although space constraints severely limit the presentation of the breadth of important evidence related to HIV transmission risks that are generally ignored, a few key points and references to review papers will be noted here. Of greatest importance for the present exposition is the extremely low risk of HIV transmission through PVI for healthy persons of reproductive age, the underestimation of the prevalence and risk associated with anal intercourse, and the underestimation of the prevalence and risk associated with punctures, especially those in medical and quasi-medical settings in sub-Saharan Africa and some other parts of the developing world [103–128].

Few studies of HIV transmission risks conducted in sub-Saharan Africa have incorporated adequate measures of unsafe punctures (and other invasive procedures) in medical settings. Those that have done so concurrently with measures of sexual behavior have found that invasive medical exposures were associated with greatly increased HIV incidence or prevalence, but measures of sexual activity (nominally PVI, as well as lack of condom use) were not associated with increased HIV incidence or prevalence [103,114,129]. Mathematical models purported to support the concept of PVI transmission have been shown to be dependent on the use of assumptions grossly inconsistent with reality [130]. The largest study on mortality in American prostitutes found that AIDS deaths occurred exclusively in those shown or inferred to be injecting drug users [54]. In a large representative American sample [131], controlling for history of homosexual contact and injection drug use revealed that for non-Hispanic black men, prevalent HIV infection was not positively associated with lifetime number of sex partners. The focus on black men is key not only because they are the most severely affected demographic (sex by race) group in the United States, but also because their base rate of HIV infection is high enough to obtain stable statistical associations.

Perhaps most importantly, when the vagaries and potential confounds of risk-factor epidemiology are bypassed, laboratory challenge of tissue under optimal conditions revealed that vaginal and cervical tissue could not become infected by exposure to HIV [132–134], but rectal tissue was readily infected under the same conditions [132,133] (in the cited studies, tissues used approximated those in natural conditions—none were chemically abraded to the point of inchoate disintegration, and viral exposure was restricted to surfaces relevant to normal organs, rather than allowing leakage from the sliced edges of the biopsies). As one research group noted: “Our data show that urogenital epithelial cells cannot be infected with NSI or SI phenotypic isolates of HIV-1” (p. 1208) [132]; this bench science observation was echoed by another research group: “HIV particles are not transmitted across the human vaginal mucosa and Langerhans cells do not increase HIV transmission” (p. 1263) [134].

In a study that examined immune function in both female prostitutes and male to female transsexuals (all HIV negative, matched for age, duration of prostitution, number of clients, and previous use of antibiotics), receptive anal intercourse was found to be associated with significantly decreased delayed-type hypersensitivity and CD4/CD8 ratios [135]. The authors concluded that receptive anal intercourse results in immunological abnormalities, and posited that the process is a result of rectal exposure to seminal alloantigens [135]. Similarly, rectal insemination (but not saline introduced into the rectum) of male rabbits led to immune suppression, including the development of immune complexes, sperm antibodies, and antibodies to peripheral blood lymphocyte antigens; additional immune effects including impaired humoral immune response to T lymphocyte-dependent antigens, keyhole limpet hemocyanin, and sheep red blood cells [136]. The results provide further evidence that receptive anal intercourse is immune suppressive, even in the absence of pathogens.

In contrast, PVI is associated with HIV-relevant immune benefits that were obliterated by condom use [137]. The authors concluded: “Unprotected sexual intercourse might result in alloimmunization stimulated by HLA antigens in seminal or cervicovaginal fluid. Mucosal alloimmunization may reduce infection by HIV-1” (p. 518).

Life Expectancy

Several longitudinal studies have found that a greater FSI predicts a longer life expectancy, but unfortunately most of these studies have not explicitly contrasted FSI with frequency of other sexual behaviors.

In a 25-year follow-up study, it was found that greater FSI predicted a lower annual death rate in men, whereas enjoyment of intercourse predicted lower mortality among women [138]. Similarly, greater FSI was associated with lower age-adjusted death rates in a 10-year study of British men that examined and excluded several possible confounding variables [139]. Compared with men reporting an FSI of at least twice weekly, men reporting an FSI of less than monthly (as well as men who declined to answer the question) had twice the death rate. The benefits associated with greater FSI were most apparent for reduced coronary heart disease mortality, which is the major cause of death in most countries of the First World. The researchers controlled for baseline coronary heart disease, as well as for social class, smoking, and blood pressure, thereby limiting to some degree the influence of reverse causality.

International comparisons suggest that national FSI estimates (as derived from surveys) are strongly associated with measures of national development, including life expectancy [140]. Of course, such ecological-level comparisons are subject not only to the usual epidemiological caveats, but to the risk of aggregation bias as well.

According to U.S. government estimates [141] men using sildenafil are considerably less likely to die of a myocardial infarction the day that they use the medication than would be expected from the population base rate of men that age dying from that major cause of death. One might infer that most of the men used the medication for intercourse rather than for masturbation, but this information was not provided. The results are not adjusted for some potentially important factors (including how healthy the men were the preceding day), and both intercourse itself and perhaps a direct cardioprotective effect of the medication might have some life-sustaining benefit. In contrast, it has been noted that during general physical exertion, men are 10 times more likely to have a myocardial infarction as when at rest [142].

Female crickets that mated repeatedly lived 32% longer than crickets that mated only once. Because the experimenters controlled the lifestyle of their subjects, the males provided only copulation (and sperm and seminal fluid) to the females, but not food or protection, thereby demonstrating that either the mating process itself or chemical components of the semen extended female life [143]. Higher on the phylogenetic scale, male rats who (for apparently genetic reasons) had a relatively high FSI lived 13% longer than those with a low or zero FSI [144]. In the latter study, administration of the Monoamine oxidase B (MAO-B) inhibitor deprenyl (also known as selegiline, it slows the degradation of dopamine and may have neuroprotective effects) increased FSI and the life span of the rats.

Tabular Summaries

Tabular summaries of the main studies reviewed herein are provided (Table 1 for psychological and psychophysiological, and Table 2 for physiological health aspects of different sexual behaviors). However, the Tables are not a substitute for the narrative expositions in the manuscript. The health benefits in the Tables are all worded such that a “+” in the last column indicates more health benefits (including decreased health risks); a “−” in the last column indicates less health benefits (including increased health risks); a “0” in the last column indicates nonsignificant effects; a combination of “0” and “−” or “+” indicates mixed results (often in different analyses); and the inclusion of a “?” indicates some mixed ambiguous results.

Table 1.  Psychological and psychophysiological health aspects of different sexual behaviors

Health aspect	Reference	Design	Participants	DV	Results
Note: These tables are a summary, but not a substitute for the narrative expositions in the manuscript. The health benefits are all worded such that a “+” in the last column indicates more health benefits (including decreased health risks), a “−” in the last column indicates less health benefits (including increased health risks), a “0” in the last column indicates nonsignificant effects, a combination of “0” and “−” or “+” indicates mixed results (often in different analyses), and the inclusion of a “?” indicates some mixed ambiguous results. M = men; W = women; DV = dependent variable; PVI = penile-vaginal intercourse; MMPI = Minnesota Multiphasic Personality Inventory.
Mental health satisfaction	[11]	Correlational (multivariate)	M, W	Satisfaction with one's mental health	PVI frequency: + Masturbation frequency: − Noncoital sex frequencies: 0
	[12]	Correlational	W	Satisfaction with one's mental health	Vaginal orgasm (vs. clitoral orgasm) ever: +
Relationship quality	[15]	Correlational	W	Perceived Relationship Quality Components (PRQC) Inventory dimensions: Satisfaction, Intimacy, Trust, Passion, and Love	PVI frequency: + Noncoital sex frequencies: 0/− Masturbation frequency: − (Love) PVI orgasmic frequency: +
	[16]	Correlational	W	Marital and sexual satisfaction	PVI frequency: + Cunnilingus frequency: 0 Masturbating male to orgasm: −
	[17]	Correlational	W (Kinsey data)	Marital happiness	PVI orgasmic frequency: +
	[11]	Correlational (multivariate)	M	Relationship satisfaction	PVI frequency: + Masturbation frequency: − Oral sex frequency: − Anal sex frequency: −
	[11]	Correlational (multivariate)	W	Relationship satisfaction	PVI frequency: + Masturbation frequency: −
Emotional awareness and integration	[21]	Correlational	W	Less alexithymia (Toronto Alexithymia Scale-20)	PVI frequency: + Masturbation frequency: 0 Noncoital sex frequencies: 0
Less psychological immaturity (less use of immature psychological defense mechanisms)	[24]	Correlational	W	Less use of immature psychological defense mechanisms (Defense Style Questionnaire-40)	Vaginal orgasm: + Clitoral orgasm: 0/− Clitoral orgasm during PVI: 0/−
	[25]	Correlational	M + W	Less use of immature psychological defense mechanisms (Defense Style Questionnaire-40)	PVI frequency without condoms: + PVI frequency with condoms: − Masturbation orgasm: − Clitoral orgasm during PVI (W): 0/− Vaginal orgasm: +
	[31]	Correlational	W	Less use of immature psychological defense mechanisms (Defense Style Questionnaire-40)	Clitoral orgasm during PVI: − Vaginal orgasm: + Masturbation orgasm: − Anal sex: − Vibrator use: −
Less depression and less suicide attempts	[34]	Correlational	M	Lower MMPI depression scores	Masturbation ejaculations: − Partnered sex ejaculations: 0
	[35]	Correlational	W	Lower Beck Depression Inventory scores	Masturbation: − Desire for masturbation: − Desire for partnered sexual activity: 0
	[36]	Correlational	W	Less likelihood of lifetime major depressive disorder (Structured Clinical Interview for the DSM−IV)	Masturbation frequency: −
	[38]	Correlational	W	Lower Beck Depression Inventory scores and less suicide attempts	PVI without condoms: + PVI with condoms: −
	[33]	Correlational (cross-sectional and longitudinal)	W	Lower Hopkins Symptom Checklist depression scores	Condom use: −
	[43]	Correlational	M + W	Less suicide contemplation Less deliberate self-harm	Homosexual (vs. Heterosexual): − Homosexual (vs. Heterosexual; M): −
	[41]	Correlational	M + W	Less mood disorders	Homosexual (vs. Heterosexual): −
Happiness	[37]	Correlational	M + W	Self-rated happiness	Masturbation: −
Less psychoticism and neuroticism	[47]	Correlational (Twin study)	M + W	Lower Eysenck Personality Questionnaire (EPQ−R) Neuroticism and Psychoticism scores	Homosexual and bisexual (vs. Heterosexual): − (Note: Effects were found to be associated with genetic but not environmental influences)
Less schizophrenia	[48]	Correlational	M + W	Less likelihood of schizophrenia diagnosis	PVI frequency: + Masturbation frequency: 0
	[49]	Correlational	M	Less likelihood of schizophrenia diagnosis	Masturbation frequency: −
	[50]	Correlational	W	Less likelihood of schizophrenia diagnosis	PVI orgasm: +
Less anorexia nervosa	[50]	Correlational	W	Less likelihood of anorexia nervosa diagnosis	PVI orgasm: +
Less neurotic disorders	[50]	Correlational	W	Less likelihood of neurotic disorder diagnosis	PVI orgasm: +
	[51]	Correlational	W	Less likelihood of neurotic disorder diagnosis	PVI orgasm: + Clitoral orgasm: 0
Less prostitution (related to various psychological disorders; [53])	[52]	Correlational	W	Less likelihood of prostitution	PVI orgasm: +
Improved erectile function	[55]	Clinical trial of intracavernosal injection (not blind)	M	Effect of restoration of erectile function	PVI frequency: + Masturbation frequency: − General psychiatric symptomatology reduction: +
Awareness and integration of vaginal sensations	[22]	Mixed correlational and experimental	W	Concordance of vaginal pulse amplitude and subjective sexual arousal	PVI orgasmic consistency: + Masturbation orgasmic consistency: 0 Noncoital orgasmic consistency: 0
	[23]	Mixed correlational and experimental	W	Concordance of vaginal pulse amplitude and subjective sexual arousal	PVI orgasmic consistency: + Masturbation orgasmic consistency: 0 Noncoital orgasmic consistency: 0
Reduced pain: Prostatodynia	[56]	Correlational	M	Reduced pain and voiding symptoms	PVI: + Masturbation: −
Reduced pain: Experimental	[57,58]	Experimental	W	Pain levels	Vaginal >> Clitoral

Table 2.  Physiological health aspects of different sexual behaviors

Health aspect	Reference	Design	Participants	DV	Results
Note: These tables are a summary, but not a substitute for the narrative expositions in the manuscript. The health benefits are all worded such that a “+” in the last column indicates more health benefits (including decreased health risks), a “−” in the last column indicates less health benefits (including increased health risks), a “0” in the last column indicates nonsignificant effects, a combination of “0” and “−” or “+” indicates mixed results (often in different analyses), and the inclusion of a “?” indicates some mixed ambiguous results. M = men; W = women; DV = dependent variable; PVI = penile-vaginal intercourse; FSH = follicle-stimulating hormone; LH = luteinizing hormone; RCT = randomized controlled trial; MTF = male-to-female; ns = nonsignificant.
Life expectancy	[138]	Correlational: Longitudinal (25 years)	M + W	Lower annual death rate	PVI frequency: + (M) Enjoyment of PVI: + (W)
	[139]	Correlational: Longitudinal (10 years)	M	Lower annual death rate	PVI frequency: +
	[140]	Correlational: International ecological	M + W	Life expectancy (national)	PVI frequency: +
	[143]	Experimental	Crickets (female)	Life expectancy	PVI frequency: +
	[144]	Mixed correlational and experimental (deprenyl trial)	Rats (male)	Life expectancy	PVI frequency: +
Vaginal function	[60]	Review	W	Vaginal tone, oxygenation, blood flow	PVI: + Condoms: − (oxygenation, blood flow effects?)
Less functional musculoskeletal disturbance	[61]	Correlational	W	Anatomically normal gait	Vaginal orgasm history: + Clitoral orgasm history: 0
Less body fatness	[62]	Correlational	M + W	Slimmer waist and/or hip circumference	PVI frequency: + Masturbation frequency: − Noncoital sex frequency: 0/+
	[63]	Correlational	Rats (Male)	Nonobese vs. obese	PVI frequency: + Cunnilingus: 0
Nutritional enhancement effects	[69]	Experimental (RCT)	M + W	Effect of 14 days 3,000 mg/day ascorbic acid vs. placebo	PVI frequency: + (effect attributable to W) Masturbation frequency: 0 Noncoital sex frequency: 0
Better heart rate variability	[71]	Correlational	M + W	Heart rate variability (resting)	PVI frequency: + Masturbation frequency: 0 Noncoital sex frequency: 0 Subjective importance of PVI: +
	[70]	Correlational	M + W	Heart rate variability (resting)	PVI frequency: + Masturbation frequency: 0 Noncoital sex frequency: 0 Subjective importance of PVI: +
Lower resting blood pressure	[71]	Correlational	M + W	Lower diastolic blood pressure (resting)	PVI frequency: + Masturbation frequency: 0 Noncoital sex frequency: 0 Note: ns results for diastolic blood pressure in [70]
	[72]	Correlational	M	Less likelihood of hypertension (newly diagnosed and untreated)	PVI frequency: +
Less blood pressure reactivity to acute stress	[73]	Mixed (correlational with sexual behavior, experimental stressor)	M + W	Less blood pressure increase before and during stress, and faster recovery after stress	PVI (past 2 weeks): + Masturbation (frequency past 2 weeks): 0/− Pattern: PVI only > PVI + days with only other sexual activities ≥ only other sexual activities
Better exercise value	[74]	Experimental	M	Better oxygen uptake, and blood pressure and heart rate stimulation	PVI > masturbation
Less pre-eclampsia risk (3 studies)	[75−77]	Correlational	W	Less pre-eclampsia risk	Condoms: − PVI frequency: + Fellatio: +
Improved testosterone levels	[78]	Experimental/Clinical	M	Restoration of normal testosterone level with various nonhormonal treatments of erectile dysfunction	PVI frequency: +
Decreased menopausal hot flashes	[81]	Experimental (randomized nonblind trial of coitus vs. no coitus)	W	Effect on menopausal hot flashes and LH and FSH normalization	PVI: +
Less intense cortisol response to stress	[82]	Mixed (correlational with sexual behavior, experimental stressor)	M + W	Less intense cortisol response to stressor	Earlier age at first PVI: + (W, ns trend for M)
Increases in dopamine levels in the nucleus accumbens	[83]	Experimental	Rats (Female)	Increases in dopamine levels in the nucleus accumbens	PVI: + Mounting/cuddling/external stimulation without PVI: 0
Greater prolactin increase following orgasm	[86]	Experimental	M + W	Greater prolactin increase following orgasm	PVI >> masturbation
Better sperm quality	[89]	Experimental	M	Volume of seminal plasma, sperm count, sperm motility, and percentage of morphologically healthy sperm	PVI > masturbation
	[91]	Experimental	M	Greater sperm volume and quality	PVI > masturbation
Better prostatic secretory function	[89]	Experimental	M	Better prostatic secretory function	PVI > masturbation
Lower prostate specific antigen levels and fewer prostate abnormalities	[88]	Correlational	M (with erectile dysfunction)	Lower prostate specific antigen levels and fewer prostate abnormalities	Masturbation: −
Better elimination of waste products	[90]	Experimental	M	Increased concentrations and total amounts of prostaglandin E and polyamines (putrescine, spermidine and spermine)	PVI > masturbation
Less prostate cancer risk	[95]	Correlational	M	Less prostate cancer risk	PVI frequency: 0 Masturbation frequency: ?
	[96]	Correlational	M	Less prostate cancer risk	PVI frequency: + Masturbation frequency: − Prostitute use: −
	[97]	Correlational	M	Less prostate cancer risk	PVI frequency: + Homosexual partners: − Prostitute use: −
Less breast cancer risk	[98]	Correlational	W	Less breast cancer risk	PVI frequency: + Condoms: − Coitus interruptus: −
	[99,100]	Correlational	W	Less breast cancer risk	PVI frequency: + Condoms: − Coitus interruptus: −
	[101]	Correlational	W	Less breast cancer risk	Lifetime number of sexual partners: +
Tissue could not become infected by exposure to HIV (3 studies)	[132–134]	Experimental (tissue challenge without excess damage to the tissue)	W (biopsies)	Tissue could not become infected by exposure to large quantities of HIV	Vaginal: + Cervical: + Rectal: − Note: See the relevant section for important extensive references to other studies with similar implications
Less impairment of immune function	[135]	Correlational	W + MTF Transsexuals	Delayed-type hypersensitivity (DTH) and CD4/CD8 ratios	Receptive anal intercourse: −
	[136]	Experimental	Rabbits (Male)	Less immune complexes, sperm antibodies, and antibodies to peripheral blood lymphocyte antigens; additional immune effects including impaired humoral immune response to T lymphocyte-dependent antigens, keyhole limpet hemocyanin, and sheep red blood cells	Rectal insemination (vs. saline introduced into the rectum): −
Immune benefits including possible resistance to HIV-1 infection	[137]	Correlational	M + W	resistance to HIV-1 infection by challenging activated CD4-positive T cells with CCR5-binding and CXCR4-binding HIV-1 strains	PVI: + Condoms for PVI: −

Discussion

The present overview indicates that there are many psychological and physiological health benefits associated with one specific sexual activity. It is specifically PVI, competently performed and sensitively experienced, that is associated with (as indicated by the correlational research designs), and produces (as indicated by the experimental research designs) aspects of better mental and physical health. This is not the case for other sexual behaviors (masturbation and anal intercourse are associated with poorer health indices, effects not attributable simply to lack of PVI). It is also not the case when PVI is done poorly, psychologically dissociated, or impaired.

The multimethod evidence presented here that specifically PVI and the orgasm it produces (rather than other sexual behaviors and responses) is associated with indices of better psychological and physiological health has several implications. Before addressing those implications, methodological issues and possible mechanisms for the observed effects are discussed.

Experimental, Correlational, and Quasi-Experimental Research Designs

There were a few experimental studies reviewed herein, and the potentially higher standard of evidence that they can provide leads to some of the experimental findings being noted here again.

In the randomized trial of coitus (vs. no coitus) for the treatment of menopausal hot flashes, PVI decreased hot flash symptoms, an effect perhaps mediated by the observed changes in LH and FSH [81]. In addition to the specific health outcome, the findings offer support for PVI leading to better hormonal regulation (and the secondary benefits thereof).

The randomized controlled trial of high dose ascorbic acid [69] indicated that better health (in the form of better nutrition) can lead to greater frequency of specifically PVI but not other partnered or solitary sexual behaviors (ascorbic acid also improved mood, so it is unclear if mood enhancement was part of the causal pathway). Thus, in addition to effects of specifically PVI leading to better health, better health can lead to greater frequency of specifically PVI, and this can be the case even in nominally healthy persons (rather than only as a consequence of identified morbidity). As noted in the Introduction, there can be bi-directional causality, as in vicious or virtuous circles of health.

In the laboratory, PVI was shown to produce greater physiological exercise value than masturbation or masturbation by a partner [74].

In the laboratory, PVI caused both sexes to manifest a postorgasmic prolactin increase 400% greater than following masturbation climax (adjusted for a control condition) [86]. The greater homeostatic dopamine modulating effects might be among the mechanisms involved in the psychological and physiological benefits associated with PVI rather than with any other sexual activities [86].

In multiple within-subjects studies [89–91] comparing semen samples produced by PVI and by masturbation, PVI led to indicators of better prostate function, better quality sperm, better semen volume, and elimination of more waste products. Although not a health outcome per se, this has implications not only for prostate health, but quite possibly for other beneficial consequences of more effective elimination of waste products.

Female crickets that were allowed to mate repeatedly lived 32% longer than crickets that mated only once. Experimental controls allowed the inference that either the mating process itself or chemical components of semen extend female life [143]. Although the research subjects are a different phylum than the readers of this review, the finding is consistent with the human life expectancy epidemiological results and with the human PVI seminal exposure findings (related to better immune function, better vaginal blood flow, and less depression) reviewed herein.

The majority of studies in this review used correlational (or risk-factor epidemiological) research designs (this problem is difficult to overcome, given the nature of some of the variables to which research subjects cannot be randomly assigned, such as immature psychological defense mechanisms). In such research designs, there is always the possibility that unmeasured third variables influence both the predictor variable and the outcome of interest (e.g., see below for a discussion of possible heritable factors).

Possible Mechanisms

The present review implied some mechanisms that might, at various explanatory levels, clarify some of the bases for the health benefits associated with specifically PVI. In addition to those processes noted in the text above, a few such mechanisms will now be briefly described. Although these possible mechanisms are examined separately, it is reasonable to conjecture that various mechanisms interact. For example, psychological outlook can modify some effects of physiological and pharmacological stimuli. The psychological outlook itself would be significantly influenced by genes (as filtered by natural and sexual selection, but also influenced by individual differences in the load of harmful mutations [145]) interacting with developmental influences (including those which support or impair psychosexual development), and would involve various neurophysiological processes (which themselves are conditioned to some degree by experience).

Possible Mechanisms: Evolutionary Issues

First, there is the distal evolutionary theoretical level: out of the universe of possible sexual activities, there is only one that is directly relevant to gene propagation. Evolutionary pressures strongly reward behaviors and mutations even slightly associated with increased likelihood of gene propagation. The difference between PVI and other sexual behaviors is not slight. The mechanisms by which such evolutionarily mandated rewards might operate range from direct physiological mechanisms (responding favorably to PVI but neutrally or unfavorably to other sexual activities) to mechanisms secondary to the evolutionary behavioral “success” of specifically PVI being rewarded by better physical and mental health (and perhaps the evolutionary behavioral failure of other sexual activities such as masturbation being punished by poorer physical and mental health). In addition to whatever skills and characteristics might be required to obtain a sexual partner, there might be additional skills and characteristics required for PVI that exceed those required for other partnered sexual activities. Evolution might also reward engaging in the one potentially reproductive behavior with various psychological (including a sense of accomplishment and/or intimacy, for some) and physiological benefits, and provide incremental rewards when there is no impediment to ejaculating into the vagina (i.e., without condoms).

Possible Mechanisms: Neurophysiological Issues

Second, at the neurophysiological level, there are differences between various sexual activities and/or areas of stimulation. As noted earlier, for both sexes, the magnitude of prolactin release following PVI orgasm is 400% greater than following masturbation orgasm (adjusted for responses to a control condition), and there are likely important psychological and psychophysiological implications of this dopaminergically relevant process [86].

Lesions to the medial preoptic hypothalamic area of adult male rhesus monkeys led to an elimination of PVI, but no change in general social behavior or frequency of masturbation [146], implying that PVI is a neurophysiologically more complex phenomenon than masturbation.

The activation of peripheral afferents also differs between some sexual activities. The pelvic nerve conducts sensory information from the vagina and regions of the cervix to the spinal cord for transmission to the brain, whereas the pudendal nerve carries sensory information from the clitoris and external perigenital skin to the spinal cord for transmission to the brain [147]. The hypogastric nerve carries sensory information from the uterus and some regions of the cervix to the spinal cord for transmission to the brain. These nerves enter the spinal cord in different places, and have other important differences that might relate to the quality of the signal and other secondary effects of information that the nerves carry [147]. Vaginal–cervical (not clitoral) stimulation results in sensory information being conveyed by the vagus nerve to the brain, bypassing the spinal cord (and allowing even women with complete spinal cord transection to orgasm from vaginal–cervical stimulation, albeit not clitoral stimulation [148]). This vagus connection is especially interesting in light of the aforementioned studies reporting that vagus-mediated greater resting heart rate variability and lesser blood pressure stress reactivity (as well as some personality processes that might be related to greater vagus activity) are associated with PVI but not other sexual activity frequencies [70,71,73]. Among other possibilities, there may be a “tuning” of autonomic functioning by habitual behaviors [149], and such tuning might influence not only resting autonomic tone, but also the response to subsequent sexual activity.

Transection of the female rat pelvic nerve reduced vaginocervical-stimulation induced Fos immunoreactivity in the medial preoptic area, bed nucleus of the stria terminalis, ventromedial hypothalamus, and medial amygdala. In contrast, transection of the clitorally linked pudendal nerve had no such effect [150]. This and other studies suggest that (vaginal) pelvic nerve but not (clitoral) pudendal nerve activity leads to crucial brain effects from PVI (including possibly facilitation of readiness for future PVI).

The hormone oxytocin has several functions, and appears to have a role in promoting pair-bonding. When oxytocin is injected into female laboratory animals, they become more sexually receptive [147]. This is the case not only in intact animals, but also in those whose (clitorally relevant) pudendal nerve has been severed. However, if the (vaginally relevant) pelvic nerve is cut, oxytocin no longer increases sexual receptivity [147]. Therefore, it is the nerve carrying sensation from PVI, and not the nerve carrying signals from external clitoral stimulation, that is required for the hormone oxytocin to manifest some of its functions.

Although there are sexual behavior differences in postorgasmic prolactin effects (and resting heart rate variability and blood pressure reactivity, suggesting Vagus nerve effects) for both sexes, the peripheral genital nerve structures differentiating intercourse from other sexual activities are not as obvious for males as for females. However, besides differentiation at the level of specific nerves, for both sexes, there is also the issue of patterns of interactive stimulation that can differentiate various sexual behaviors. A few studies have provided some technical details regarding genital pulsatile communication. Penile thrusting in the vagina and against the cervix results in the vaginal muscles gripping the lower part of the penis [151,152], potentially leading to not only identification of the process being PVI as distinguished from other sexual behaviors (including stimulation with nongenital objects), but also a virtuous circle of genital-to-genital response.

As suggested to varying degrees in various studies above [25,38,60,137,143], there is also the possibility that the absorption of semen by the vagina and absorption of vaginal secretions by the penis activate psychological and physiological health-supporting pathways.

There are likely other physiological (and perhaps implicit psychological) mechanisms by which the organism differentiates PVI from other sexual behaviors (to induce better health). One need only pause to reflect on the evolutionary value of differentiating perhaps one million different colors as compared with the evolutionary value of mechanisms for differentiating between activities that could or could not perpetuate the “selfish gene”[153].

Possible Mechanisms: Psychosexual Issues

Third, there are the psychosexual developmental issues. As conjectured by Freud, problems in psychosexual development can produce chronological adults with psychological (including psychophysiological) problems, and associated impairment of the ability to appreciate fully PVI. Several studies described herein, including those on alexithymia and on immature psychological defense mechanisms [21,24,25,31], speak to those issues. The results of those studies could generally be understood as an example of poorer health leading to avoidance of or impairment of specifically PVI (including choosing to use condoms for PVI, lack of vaginal orgasm, or choosing other sexual activities in lieu of PVI), but one cannot not rule out that there might be a role for impairment or avoidance of PVI leading to increased use of immature defense mechanisms.

Although the concept of impaired psychosexual development leading to avoidance of specifically PVI frequency or response might seem arcane for some readers, some basic aspects of the process can be operationalized in an obvious manner. For example, in a laboratory study [154], adult male hamsters underwent a single-trial learned aversion to vaginal secretion. Subsequently, they had a lower PVI frequency than did the control group. Although the hamsters with the learned aversion to PVI also had a lower frequency of grooming their own genitals, they did not differ in the amount of time they spent licking vaginas [154]. Thus, the unpleasant early experience did not even lead to avoidance of vaginal secretions, but to avoidance of the deeper function of PVI.

Engaging in masturbation or even partnered sexual behavior in the absence of PVI on some days appears to detract from the psychological [11] and physiological [73] benefits associated with having PVI on other days. Among the possible mechanisms for this is a psychological variant on the autonomic tuning process [149] described above: PVI might be approached in essentially the same spirit (including less relatedness, which is manifestly the case for masturbation) as the other sexual activities, leading to less psychological and physiological benefit from PVI. Another possibility [73] is that PVI is being avoided on some but not all occasions (by substituting other sexual activities), which might be because of a subtler psychosexual dysregulation than complete avoidance of PVI.

Possible Mechanisms: Genetic Factors

Fourth, genetic factors could conceivably produce both greater FSI (but not greater frequency of other sexual behaviors) and better physical and mental health. Suggestive, if weak, support for this hypothesis might be gleaned from the finding that male rats who—for apparently genetic reasons—had a relatively high FSI lived 13% longer than those with a low or zero FSI [144]. Studies of women's intercourse orgasm consistency (unfortunately, not specifying vaginal orgasm in any of the three studies) reported a heritability of 31–34% for intercourse orgasm consistency [155,156] (and 45–51% heritability estimates for masturbation orgasmic consistency), as well as significant associations with extraversion, being open to new experience, and less neuroticism [157]. These dimensions of personality also have a significant heritability. The association of coital orgasm consistency with less neuroticism would be consistent [158] with the findings of immature defense mechanisms being associated with lesser vaginal orgasm consistency [24,25,31].

When feasible, future research on sexual behavior and sexual medicine might incorporate sophisticated quasi-experimental designs that consider the effect of heritable factors. When one such behavior genetic quasi-experimental design was used, it was revealed that earlier age at first intercourse is associated with less (rather than as usually assumed, more) subsequent conduct disorder in American samples [159].

Implications for Clinical Assessment, Treatment, Research, and Sex Education

Clinicians and researchers need to be specific in examining various sexual behaviors (including details of how PVI might have been modified from its pure form, such as condom use or clitoral masturbation during PVI; studies noted herein found these practices to be associated with poorer functioning than undisturbed PVI). Despite the scientific and clinical importance of being specific regarding the differences between specific sexual behaviors, some of the most commonly used research and clinical measures of sexual function explicitly obscure the differences between various sexual behaviors, thereby limiting their utility and ability to help patients and scientific inquiry. For example, Female Sexual Function Index [160] questions refer to “sexual activity or intercourse,” thereby losing the essential differentiation (the questions that specify “vaginal penetration” are those dealing with pain). The International Index of Erectile Function [161] includes many questions with the nonspecific “sexual activity includes intercourse, caressing, foreplay & masturbation,” and some with “sexual intercourse is defined as sexual penetration of your partner” (which could involve anal intercourse, and perhaps even broader interpretations might be made by some people). More specific questionnaires specifying more specific sexual behaviors [24] would be far more informative, and allow sexual medicine to advance.

In addition, sex education should begin to be honest regarding physiological and psychological health differences between specific sexual behaviors. In a recent large representative survey of Czech women, vaginally orgasmic women were significantly more likely than vaginally anorgasmic women to report having being told in childhood or adolescence that the vagina was an important zone for inducing female orgasm [162].

Treatment (including primary prevention) should be supportive of PVI and the orgasm it produces, and not assume that behavioral treatment of sexual dysfunction must involve masturbation as part of the process. There are effective PVI-based treatments for premature ejaculation [163] and for female orgasmic dysfunction [164]. It has been noted that for some women, repeated orgasm from clitoral stimulation can interfere with the development of pathways leading to vaginal orgasm [165,166]. In a recent large representative survey of Czech women, risk of female sexual arousal disorder with distress (albeit not without distress) was much lower for women who had a history of vaginal orgasm [167]. The concept of sexual dysfunction merits both broadening and more specificity (e.g., lack of vaginal orgasm [31]) in light of the research described herein.

Among the factors shown to be associated with greater likelihood of women's orgasm with a partner is duration of PVI, but not duration of foreplay (in contrast to the assumptions of many practitioners) [162,168]. Future studies might examine in greater detail which physiological and psychological factors lead to people avoiding (in behavior or full feeling and response) the key evolutionarily driven sexual behavior, including choosing to substitute something else for one set of genitals in lieu of genital–genital (penile–vaginal) intercourse.

There might also be implications for future research involving pharmacotherapy. Women's positive response to phosphodiesterase inhibitor treatment for sexual dysfunction was better in women with “stable and/or happy relationships” (p. 156), and nonresponse was more common in the “presence of psychological co-morbidities” (p. 156) [169]. Studies might examine the degree to which phosphodiesterase inhibitor benefit is more likely in women [170] who are more aware of their vaginal response (especially those with a history of vaginal orgasm).

Some studies of risk factors for sexual dysfunction might suggest future studies on the possible influence of better sexual function on the prevention of the “risk factors.” For example, a global measure of women's sexual dysfunction found multivariate associations with lower high density lipoprotein cholesterol, and higher low density lipoprotein cholesterol, triglycerides, body mass index, and age [171]. Given that more favorable lipid profiles may be associated with more mature emotion expression [172,173], and that specifically PVI frequency and response is associated with both slimness [62] and better emotional functioning [11,12,21,24,25], future research might examine the possibility that unfavorable lipid profiles might be secondary to sexual behavior factors. Various studies have indicated associations between depression, cardiovascular disease or risk, and sexual dysfunction (especially erectile dysfunction). In addition to such obvious pathways as cardiovascular disease leading to erectile dysfunction leading to depression, other pathways are also likely [174].

Conclusion

Based upon a broad range of methods, samples, and measures, the research findings are remarkably consistent in demonstrating that one sexual activity (PVI and the orgasmic response to it) is associated with, and in some cases, causes processes associated with better psychological and physical functioning. Other sexual behaviors (including when PVI is impaired, as with condoms or distraction away from the penile–vaginal sensations) are unassociated, or in some cases (such as masturbation and anal intercourse) inversely associated with better psychological and physical functioning.

Sexual medicine, sex education, sex therapy, and sex research should disseminate details of the health benefits of specifically PVI, and also become much more specific in their respective assessment and intervention practices.

Conflict of Interest: None. However, Dr. Brody discloses that he was or is a consultant for Bayer Schering.

Statement of Authorship

Category 1

• (a) 
  Conception and Design
  Stuart Brody
• (b) 
  Acquisition of Data
  Stuart Brody
• (c) 
  Analysis and Interpretation of Data
  Stuart Brody

Category 2

• (a) 
  Drafting the Article
  Stuart Brody
• (b) 
  Revising It for Intellectual Content
  Stuart Brody

Category 3

• (a) 
  Final Approval of the Completed Article
  Stuart Brody