CHAPTER 61 Chronic pelvic pain
Biology of Pain
Pain is defined by the International Association for the Study of Pain as ‘an unpleasant sensory and emotional experience associated with actual or potential tissue damage, or described in terms of such damage’ (IASP Task Force on Taxonomy 1994). The emphasis in this definition is that pain is an experience rather than a neurophysiological process, and is therefore subjective. There is no such thing as ‘objective’ pain. This does not mean, however, that pain cannot be studied or that reproducible experimental models cannot be devised. Neuroscience has advanced knowledge of the biological processes underlying pain. It is no longer appropriate to discuss the subject in terms of ‘pain pathways’, by analogy with a wiring diagram, or even to consider that the nervous system conveys perceptions such as touch or pain. Rather, the elements of the nervous system, comprising the ‘different fibres, tracts, pathways and nuclei process and convey information about bodily stimulus events’ (Berkley and Hubscher 1995). In this model, pain is a central nervous system construct derived from the totality of sensory input rather than from the activation of a particular pathway. Even the anatomy of neural pathways in the adult cannot be considered immutable; plasticity of the central nervous system has been clearly demonstrated in reproducible animal experimental models, such that a nerve injury leads to sprouting of afferent fibres within the dorsal horn of the spinal cord (Woolf et al 1992).
Special features of chronic and visceral pain
The processes underlying the transition from a single or repeated acute painful episode to a chronic pain state are of great potential clinical significance. Patients often relate the onset of their chronic condition to an event such as an acute infection or surgical procedure. Moreover, visceral sensory mechanisms differ in certain respects from those found in cutaneous tissues. Key mechanisms potentially underlying the transition from acute to chronic visceral pain states are the activation of silent afferents in viscera (McMahon 1997) and central sensitization (Woolf 1983). There are certain similarities but also important differences in the context of visceral sensation between central sensitization and wind-up, where the response to repeated C-fibre stimulation is progressively augmented (Herrero et al 2000). It had been hoped that the development of central sensitization and/or wind-up could be modulated by analgesics or other agents using a therapeutic strategy of pre-emptive analgesia (Dickenson 1997), although clinical applications of this concept have been disappointing.
Vascular pain and pelvic congestion
By analogy with the pathogenesis of cerebral migraine, it has been suggested that pain associated with pelvic congestion might arise through vascular perturbation at the ovarian or uterine arteriolar level, resulting in the release of endothelial factors such as ATP which act by exciting sensory nerves in the outer muscle coat of venules and veins, and which also cause vasodilatation through the release of nitric oxide (Stones 2000).
Sex differences in pain
There is a marked sex difference in the prevalence of a number of chronic painful conditions unrelated to the reproductive tract in the general population, such as irritable bowel syndrome (IBS), temporomandibular dysfunction and interstitial cystitis, which has prompted research on the potential underlying mechanisms using animal models such as bladder irritation (Bon et al 1997), and uterine and vaginal distension (Bradshaw et al 1999). Illustrating the complexity of the central mechanisms underlying nociception, vaginal hyperalgesia resulting from induced endometriosis in rats was exacerbated in the presence of increased oestradiol levels, whereas in ovariectomized animals, hyperalgesia was relieved by oestradiol (Berkley et al 2007).
Studies of human volunteers have investigated pain thresholds in women at different stages of the menstrual cycle using electrical stimulation, pressure dolorimetry on the skin, thermal stimulation and induced ischaemic pain. Patterns of variation in response seen in studies using these modalities were different, and a meta-analytical review of pain perception across the menstrual cycle (Riley et al 1999) estimated the effect size for menstrual cycle fluctuation of pain sensitivity between the most and least sensitive phase to be 0.40. The effect size for sex difference was approximately 0.55, indicating that hormone variability could account for a substantial proportion but not all of the observed differences.
Illustrating some of the possible underlying mechanisms, a study of the discrimination of thermal pain in male and female volunteers showed that women had a lower pain threshold and tolerance, as also noted in a number of other studies (Fillingim et al 1998). A further complexity to be considered when interpreting studies of human volunteers is that subjects may well have prior pain experience which could influence responses to experimental pain (Fillingim et al 1999). Finally, the potential influence on pain perception of exogenous hormone therapy needs to be considered. In 87 postmenopausal women consulting for chronic orofacial pain, hormone replacement therapy (HRT) was associated with greater reported levels of pain, with substantial effect sizes between 0.39 and 0.62 (Wise et al 2000). The possibility of increased pain as a result of HRT use has not been addressed prospectively but may be clinically important, especially in women with vulval pain where oestrogen deficiency changes are thought to coexist.
Genetic variation in susceptibility to pain
There is now evidence for a genetic basis for the variation in response to pain seen in human experimental studies. Shabalina et al (2009) studied polymorphisms of the ‘mu’ opioid receptor OPRM1 gene locus, and reported a strong association between the single nucleotide polymorphism rs563649 and individual variations in pain perception. There are also relevant observations from clinical populations with gynaecological pain; for example, a twin study showed that while 39% of the variance in reported menstrual flow was accounted for by genetic factors, the corresponding figure for dysmenorrhoea was 55% and for functional limitation from menstrual symptoms was 77% (Treloar et al 1998).
Epidemiology of Pelvic Pain: Populations, Consulting and Referral Patterns
The first assessment of the prevalence and economic burden associated with chronic pelvic pain was based on extrapolation from hospital practice in the UK, and estimated the prevalence in women at 24.4 per 1000. The annual direct treatment cost was estimated at £158 million, with indirect costs of £24 million (Davies et al 1992). The above prevalence now appears an underestimate since data from population-based studies have become available. A telephone survey was undertaken in the USA using a robust sampling methodology (Mathias et al 1996). Women aged 18–50 years were interviewed about pelvic-pain-related symptoms. In total, 17,927 households were contacted, 5325 women agreed to participate and 925 reported pelvic pain of at least 6 months’ duration, including pain within the past 3 months. Following exclusion of pregnant and postmenopausal women and those with pain that was solely cycle related, 773/5263 (14.7%) were identified as suffering from chronic pelvic pain. Direct costs of health care, estimated from Medicare tariffs and hence very conservative, were $881.5 million, patients’ out-of-pocket expenses were estimated at $1.9 billion and indirect costs due to time off work were estimated at $555.3 million.
In the UK, the population perspective was provided by a postal survey of 2016 women selected at random from the Oxfordshire Health Authority register of 141,400 women aged 18–49 years (Zondervan et al 2001b). Chronic pelvic pain was defined as recurrent pain of at least 6 months’ duration, unrelated to periods, intercourse or pregnancy. For the survey, a ‘case’ was defined as a women with chronic pelvic pain in the previous 3 months, and on this basis the prevalence was 483/2016 (24.0%). Among those with pelvic pain, dysmenorrhoea was reported by 81% of those who had periods, and dyspareunia was reported by 41% of those who were sexually active. Among women who did not have chronic pelvic pain as defined above, dysmenorrhoea was reported by 58% of those who had periods, and dyspareunia was reported by 14% of those who were sexually active.
An estimate of the consulting pattern associated with pelvic pain was obtained using a national database study of UK general practices (Zondervan et al 1999b). Data relating to 284,162 women aged 12–70 years who had a general practice contact in 1991 were analysed to identify subsequent contacts over the following 5 years. The monthly prevalence rate was 21.5/1000 and the monthly incidence rate was 1.58/1000. These prevalence rates are comparable with those for migraine, back pain and asthma in primary care. Older women had higher monthly prevalence rates; for example, the rate was 18.2/1000 in women aged 15–20 years and 27.6/1000 in women over 60 years of age. This association is thought to be due to persistence of symptoms in older women, with the median duration of symptoms being 13.7 months in 13–20 year olds and 20.2 months in women over 60 years of age (Zondervan et al 1999a). It is clear that future population-based studies need to include older women.
Among 483 women with chronic pelvic pain participating in the Oxfordshire population study discussed above, 195 (40.4%) had not sought a medical consultation, 127 (26.3%) reported a past consultation and 139 (28.8%) reported a recent consultation for pain (Zondervan et al 2001b). Of those women identified as cases of pelvic pain in the national general practice database, 28% were not given a specific diagnosis and 60% were not referred to hospital (Zondervan et al 1999a). The US population-based study discussed above also drew attention to the large numbers of women who have troublesome symptoms but do not seek medical attention: 75% of this sample had not seen a healthcare provider in the previous 3 months. It might be thought that not seeking care would be an indicator of milder symptoms; indeed, in the US study, those who did seek medical attention had higher pain and lower general health scores than those who did not seek medical attention. However, among those not seeking help, scores for pain and functional impairment were still substantial. Lack of use of medical services might reflect sociocultural factors, but could equally reflect previous unsatisfactory experiences of investigation and treatment, as discussed later in this chapter.
Among women referred to a UK gynaecology outpatient clinic with chronic pelvic pain of at least 6 months’ duration, the impact of the condition on quality of life was assessed using the Short Form-36 (SF-36) questionnaire (Stones et al 2000). Figure 61.1 shows a comparison of physical and mental component summary scores derived from the eight SF-36 subscales (Jenkinson et al 1996a) in this hospital population with the data from the Oxfordshire population study described above. These summary scores are adjusted such that 50 represents a normal population mean. It will be noted that, as might be expected, the hospital sample is consistent with the trend, evident in those who have previously or recently sought medical advice compared with those who have not, to greater impairment of function on both physical and mental components.
Figure 61.1 Means ± 95% confidence interval for the Short Form-36 mental component and physical component summary scales. Women with pelvic pain (a) seen in a general gynaecology clinic (Stones et al 2000), or in a postal survey (Zondervan et al 2001b) reporting (b) a recent consultation, (c) a previous consultation or (d) no medical consultation. Arbitrary units such that 50 represents the mean of a normal population.
From Stones RW, Selfe SA, Fransman S, Horn SA (2000) Psychosocial and economic impact of chronic pelvic pain. Baillière’s Clinical Obstetrics and Gynaecology 14:27–43 and From Zondervan KT, Yudkin PL, Vessey MP et al (2001b) The community prevalence of chronic pelvic pain in women and associated illness behaviour. British Journal of General Practice 51(468):541–547.
Psychological Factors in Chronic Pelvic Pain
Many writers on pelvic pain have attempted to characterize an adverse psychological profile which might help the clinician to distinguish between ‘organic’ and ‘psychogenic’ pain; a distinction not in keeping with current neurophysiological understanding as discussed above. It is likely that some of the psychological disturbance that can be identified in women with chronic pelvic pain is the result of longstanding pain symptoms and unsatisfactory treatment, rather than the cause of pain (Slocumb et al 1989). Pain may contribute to or confirm a sense of helplessness or a tendency to engage in catastrophic thoughts, and may itself be exacerbated by them (Horn and Munafo 1997). On the other hand, it is clear that concomitant mood disturbance or specific psychopathology may impair the patient’s ability to cope with her symptoms and contribute to functional impairment, and recognition is important but often neglected in the hurried setting of a gynaecology clinic (Yonkers and Chantilis 1995). The presence of adverse psychological propensities such as catastrophizing is a factor affecting outcome of treatment for chronic pelvic pain (Weijenborg et al 2008).
Depression
Women presenting with multiple pain symptoms are at especially high risk for current mood disturbance; the likelihood of an associated mood disorder was increased six-fold in individuals with two pain complaints and eight-fold in those with three complaints (Dworkin et al 1990). In keeping with the irrelevance of the organic/psychogenic distinction for diagnosis, the absence of laparoscopically visible pathology was not associated with a higher probability of depression (Peveler et al 1995, Waller and Shaw 1995). In these studies, no differences in mood-related symptoms were identified in women with chronic pelvic pain with or without endometriosis. Antidepressant therapy may be indicated in order to alleviate depression, but sertraline was not effective for pelvic pain in a recent small but well-conducted randomized trial (Engel et al 1998).
Abuse and somatization
Child sexual or physical abuse may be an antecedent for chronic pelvic pain, but many individuals have suffered such abuse without this or other consequences in later life, and the research literature is beset with the problem of appropriate comparison groups. A study from a tertiary referral multidisciplinary clinic setting reported on the full assessment of psychological factors in three groups of 30 women with chronic pelvic pain, chronic pain of other types, or without pain identified from general practitioner records. Twelve (40%) of those with chronic pelvic pain reported sexual abuse, compared with five (17%) in each of the two comparison groups. Experience of physical violence was similar in the three groups, but women with chronic pelvic pain had higher scores for somatization, meaning the experience and communication of distress and physical symptoms without clear underlying pathology. In women with pelvic pain, abuse histories were evenly distributed among those with and without identified pelvic pathology such as endometriosis, but somatization scores were higher among those with identified pathology (Collett et al 1998). It has been suggested that the potential link between sexual abuse and pelvic pain might be that abuse is an observable marker for childhood neglect in general (Fry et al 1997), and this might explain the association in some studies with physical rather than sexual abuse (Rapkin et al 1990). With regard to mediators between abuse and pain experience, those who reported abuse were more likely to be distressed and anxious (Poleshuck et al 2005).
Identification of patients with features suggesting somatization disorder is important as failure to do so will lead to further inappropriate investigation, and treatment directed towards physical symptoms which are, in fact, manifestations of psychological distress. There are limited data regarding the prevalence of frank somatization disorder in clinic populations; in a study comparing the medical assessment with a standardized questionnaire, doctors identified 19% of patients as being potential somatizers, and approximately 5% met questionnaire criteria for overt somatization disorder (Peveler et al 1997). With regard to patients within the pelvic pain spectrum of conditions, the finding of Zolnoun et al (2008) of higher scores for somatization among patients with primary vulval vestibulitis compared with those with secondary vulval vestibulitis indicates a need for more detailed characterization of this propensity among the different clinical subpopulations.
Sociocultural Factors in Chronic Pelvic Pain
It could be argued that in life-threatening medical conditions or conditions where a technical solution gives excellent results, the patient’s experience of care and the quality of doctor–patient communication are not critical to the outcome. In contrast, in ill-defined, chronic conditions where a technical solution is unlikely to provide relief, sociocultural factors have a major influence on the patient’s decision to seek care and the referral pathway subsequently followed. The population-based studies cited above provide an indication of the size of the iceberg of symptoms below the waterline of care seeking. Similarly, in general practice, many women remain undiagnosed and unreferred. This is not necessarily inappropriate but the determinants of presentation and referral are unlikely to be disease factors alone. Once seen in a gynaecology clinic, the interaction between disease states and the consultation setting can be identified in statistical models; the presence of endometriosis was a factor predicting continuing pain 6 months after an initial hospital outpatient consultation, but so also were the patient’s initial report of pain interfering with exercise, her rating of the initial medical consultation as less satisfactory (Figure 61.2) and the individual clinician undertaking the consultation (but not the doctor’s grade or gender) (Selfe et al 1998a). The ability to establish a therapeutic rapport and meet the patient’s expectations during consultations, rather than simply convey information, may be factors that favour a good outcome (Stones et al 2006).
Often, problems are encountered not just with a particular individual clinician or aspect of care, but tend to accumulate for some patients as they move through the system. This seems to be more of a problem for women of low socioeconomic status (Grace 1995). A solution to this mismatch between women’s needs and the structures through which care is provided may be the multidisciplinary approach, which at least works against the crude compartmentalized organic/psychological causation model, although it has been suggested that a more fundamental reconsideration of the medical paradigm is required (Grace 1998). Meanwhile, within the practical constraints of time, lack of continuity of care in the hospital setting and very limited access to multidisciplinary resources, gynaecologists can at least recognize the potential impact of their own attitudes and communication styles (Selfe et al 1998b).
