Trends in the aetiology of urogenital fistula: a case of 'retrogressive evolution'?

How might one obtain further information to confirm or refute these impressions? Throughout history the establishment of causality in medicine has proved elusive [ 45 , 46 ], and the confirmation of secular trends in causality presents even greater challenges. Difficult as these issues have been to establish in the context of infectious diseases [ 47 ], they take on a completely different magnitude with a problem as complex as obstetric fistula, given its physical, biosocial, cultural, economic and geopolitical contributors. Data collection is clearly important, but even such powerful tools as the Demographic and Health Surveys Program can provide only a small proportion of the information needed to fully investigate these issues [ 48 , 49 ]. Almost all of the United Nations Millennium Development Goals potentially impact on the causation and management of fistula [ 50 , 51 ], and yet there remain substantial gaps in our achievement of these goals [ 52 ]. Given a worldwide fistula prevalence of up to 3,500,000 [ 7 ], and an incidence of up to 1 in 800 births [ 16 ], there is clearly much to be done before we can claim to be effectively managing and preventing this condition.

In a study using national Hospital Episode Statistics (HES) in England, although overall the numbers were small in global terms, Cromwell and Hilton found a 37 % increase in the number of repair or diversion operations carried out for lower urinary tract fistulae between 2002 and 2009. Considering only primary fistula repair procedures, they found a 68 % increase over the same period (from 62 primary repairs in 2002 to 104 in 2009) [ 42 ]. In a further study the same authors found that the number of hysterectomies undertaken in England over a similar period for selected benign and malignant indications fell by 12 % (from 43,014 in 2000 to 37,923 in 2008) [ 41 ]. They went on to examine the risk of VVF and urethrovaginal fistula following hysterectomy in the National Health Service in England, and found this to vary both with the indication for surgery and type of procedure undertaken, but overall increased by almost 50 % over the course of the decade studied, from 0.15 % (1 in 681 hysterectomies) in 2000 – 2002 to 0.22 % (1 in 465 hysterectomies) in 2006 – 2008 [ 41 ]. In a later study of ureteric injury and ureterovaginal fistula associated with hysterectomy the same group found the risk more than doubling over time, from 0.29 % (1 in 345 hysterectomies) in 2001 – 2005 to 0.66 % (1 in 142 hysterectomies) in 2006 – 2010 [ 43 ]. A similar trend over time in post-hysterectomy VVF has recently been reported in the USA [ 44 ].

There are several possible factors that might contribute to the findings in relation to iatrogenic (post-surgical) fistulae in the UK. The reduction in the number of hysterectomy procedures undertaken may lead to a relative increase in the number of more complex procedures being carried out, with a higher risk of operative morbidity. However, Hilton and Cromwell did not find the rate of fistula to be significantly higher in women undergoing an abdominal hysterectomy for fibroids or endometriosis than in those with menstrual problems, making this explanation unlikely [ 41 ]. The reduction in gynaecological surgical activity overall, with more widespread use of non-surgical treatments, particularly for menorrhagia, the reduced length of trainees’ working week consequent upon the European Working Time Directive, and reduced years in training, all mean that trainees, in the UK at least, obtain less surgical experience now than previously, and may therefore embark on independent practice less prepared than a decade ago.

Chassar Moir reviewed the position of VVF in Britain in 1973 [ 39 ], and Lee et al. similarly in the USA in 1988 [ 40 ]. The largest recent UK case series reflects supraregional tertiary care practices, reported by Hilton in 2012 (348 cases) [ 38 ]. In the latter series, two-thirds of cases were of surgical aetiology, with the remainder evenly distributed between ‘obstetric’, radiation-induced and miscellaneous or traumatic causes. Although any surgery carried out in the pelvis can be complicated by the development of fistula (Hillary et al., submitted for publication) [ 38 ], reports from high-income countries have shown that 70 % of surgical fistulae and 50 % of all lower urinary tract fistulae follow hysterectomy [ 41 ]. Less than 20 % of fistulae in high-income countries follow childbirth, and most of these follow obstetric interventions (caesarean section, caesarean hysterectomy, instrumental vaginal delivery, and symphysiotomy), rather than obstructed labour, and should be looked on as being iatrogenic in nature.

There is a growing perception that with more widespread understanding of the causes of obstetric fistulae in low-income and middle-income countries, and increasing rates of caesarean section, there is an increase in VVF and ureterovaginal fistulae that may be at least in part iatrogenic in nature. There is evidence to support the possibility of an increase in the incidence of iatrogenic (post-hysterectomy) fistulae in high-income countries, which coincides with a decrease in the number of hysterectomies carried out. In the first case this may reflect the amount of training, supervision and support provided to trainees; in the second it may reflect the amount of surgical experience accrued in training and workload in independent practice. In both situations it behoves governments and educational authorities to invest in training and workforce planning, and to facilitate audit and operational research, so that an appropriately trained and supervised medical workforce is maintained in the correct working environment. Alternatively, so that the right people do the right thing at the right time and in the right place. Only then can our surgical evolution in this area undergo the quantum change required to move from its present retrogressive position to become progressive. Plus ça change, plus ça peut s’améliorer – after J.-B.A. Karr (and G.W. Leibniz) Plus ça change, plus ça peut s’améliorer – after J.-B.A. Karr (and G.W. Leibniz)

Plus ça change, plus ça empire – after J.-B.A. Karr (and ‘Voltaire’) Plus ça change, plus ça empire – after J.-B.A. Karr (and ‘Voltaire’) The term ‘retrogressive’ is defined as: characterized by retrogression: as a: declining from a better to a worse state b: passing from a higher to a lower level of organization [ 1 ] characterized by retrogression: as a: declining from a better to a worse state b: passing from a higher to a lower level of organization [ 1 ] More specifically, ‘retrogressive evolution’, is seen as: A developmental trend in evolution resulting in simplification of an organism, usually through the complete or partial loss of one or more structures [ 2 ] A developmental trend in evolution resulting in simplification of an organism, usually through the complete or partial loss of one or more structures [ 2 ] Charles Darwin in his book ‘ On the Origin of Species ’ in 1859 proposed the theory of natural selection to explain the progressive advance of evolutionary processes [ 3 ]. In contrast, in his 1868 work ‘ On the Origin of Genera ’, E.D. Cope proposed that evolution may proceed not only on lines of acceleration but also of retardation [ 4 ]. That is, that evolution had been not only progressive, but at times retrogressive [ 5 ]. Huxley in his collected essays entitled ‘ The Struggle for Existence in Human Society ’ in 1888 similarly emphasised: ‘It is an error to imagine that evolution signifies a constant tendency to increased perfection. … Retrogressive is as practicable as progressive metamorphosis’ [ 6 ]. What, then, can we say about the evolution of our surgical practices in relation to urogenital fistula? It is conventional wisdom that urogenital fistulae in low-income and middle-income countries are almost exclusively of obstetric aetiology, related to prolonged neglected obstructed labour, whereas those seen in high-income countries are largely iatrogenic in nature. Furthermore, the pattern of disease reported from low-income and middle-income countries in recent years apparently reflects that seen in high-income countries over a century ago. One might anticipate that this would augur well for rapid progress towards the eradication of fistula by effective primary and secondary prevention in low-resourced environments, and the thought that retrogressive evolutionary processes might be at work in our surgical development is anathema; emerging evidence, however, suggests that some consideration of this possibility is justified.

Incontrovertibly, a woman neglected in obstructed labour for several days, ultimately delivering a stillborn infant vaginally, then surviving to develop a vesicovaginal fistula (VVF), should be described as having a ‘true obstetric’ or ischaemic fistula. In contrast, a woman identified as being in obstructed labour who is transferred expeditiously to a healthcare facility for timely caesarean section, delivering a live infant, who then develops a vesicocervical fistula, would be reasonably described as having an iatrogenic injury. But what of those women who are transferred only after considerable delay, or who undergo caesarean section for a dead baby and then develop a VVF? Or those who undergo caesarean section but then develop an urethrovaginal fistula or a midvaginal VVF or a ureterovaginal fistula? Should these lesions be looked on as ‘obstetric’, iatrogenic, or a combination? In Tanzania it is reported that 85 % of patients developing fistula now deliver in hospital, usually by caesarean section, with an increasing rate of vesicocervical and vesicouterine (probably iatrogenic) fistulae. Onsrud et al. , in the Democratic Republic of Congo (DRC), have also reported a higher rate of uterine or cervical involvement, with less surrounding fibrosis, and less apparent treatment delay in post-caesarean section fistulae [ 32 ]. They consider these to be a distinct clinical (iatrogenic) entity and highlighted the training issues discussed above. Loposso et al. similarly highlighted the risks of fistula in DRC despite the availability of caesarean section, but attributed this to the duration of obstructed labour prior to hospitalisation, and the delay in achieving caesarean delivery [ 33 ]. The various colleagues with whom the author has discussed this issue (see Acknowledgments), all with much greater experience of working in different low-income and middle-income countries in Africa and Asia, have given a range of views on the point. Hence, it is likely in different circumstances fistulae following caesarean section may have varying pathophysiology, some being wholly ischaemic, some wholly traumatic, but most with a combination of mechanisms. Similar arguments may be framed around post-caesarean section ureterovaginal fistulae. Although these are commonly looked on as being exclusively iatrogenic in aetiology, operating on, or close to, the ureter within an ischaemic field defect inevitably puts it at greater risk, and means that ureterovaginal fistula developing after caesarean section for obstructed labour is likely to be at least in part ischaemic in nature. With the increased use of caesarean section, and consequent increase in ‘repeat’ procedures, the rate of placenta praevia and associated abnormal placental implantation will continue to rise [ 34 – 36 ]. Given the technical difficulties of surgery in these circumstances, we might also anticipate a progressive increase in the rate of fistula following caesarean section. Although the association has been reported [ 37 ], a trend is not yet evident, and in the author’s own series from the UK, fistula following caesarean section or trial of attempted vaginal birth after caesarean have consistently represented 6 – 8 % of cases referred over the last three decades [ 38 ].