Williams and Kapila: Complications of Circumcision

Correspondence to: Mr. N. Williams, Department of Surgery, Hope Hospital Salford, M6 8HD, UK

Note:

The incidence of circumcision in the United Kingdom has dropped dramatically in the decade since the date of publication of this article. The most recent reported annual circumcision rate was 12,200 procedures. See  Towards evidence based circumcision of English boys: survey of trends in practice for more information.

Circumcision is probably one of the oldest of all surgical procedures. Although the origin of the practice is not entirely clear, it almost certainly began as a religious rite. That it was practised by the Egyptians is evident as the earliest mummies were found to be circumcised. In the Jewish community circumcision remains a religious ritual and is usually performed on the child's eighth day of life by a Mohel. Religious circumcision is also practiced by Muslims; the procedure is performed between the ages of 4 and 13 years. Curiously the Koran contains no specific ordinance on this subject. Currently approximately one-sixth of the world's male population is circumcised mostly on religious grounds. In Western society circumcision is performed for medical reasons, the commonest of which is phimosis^1-3. There is however, enormous variation between the circumcision rate in the UK (5-6 per cent^1,4) and that in the USA (80-90 per cent^5-7). This large discrepancy exists despite recommendations from both the British Medical Association^8,9 and the American Association (sic) of Pediatricians¹⁰ that circumcision should be performed only for medical reasons.

The aim of circumcision is to excise sufficient foreskin (both penile shaft and inner preputial epithelium) to leave the glans uncovered. As an alternative to circumcision some advocate the technique of 'preputial plasty' where a longitudinal incision of the phimosis constricting band is followed by a transverse suture¹¹. Apparently more popular in Europe, it is a method that has yet to find favour among surgeons in the UK. There are many types of circumcision but they can be broadly classified into four types: dorsal slit, shield, clamp and excision¹². All of these techniques have their strengths and their limitations together with their protagonists and critics. More often than not, however, complications arise as a result of operator inexperience rather than the method employed. All techniques aim to provide the best cosmetic result together with the lowest possible morbidity rate; in this respect the key factors to be observed are attention to asepsis, adequate but not excessive excision of the inner and outer preputial layers, haemostatis and cosmesis¹². Some authors have reported a complication rate as low as 0.06 per cent¹³, while at the other extreme rates of up to 55 per cent¹⁴ have been quoted. This reflects the differing and varying diagnostic criteria employed; a realistic figure is 2-10 per cent^3,12,15. Although haemorrhage and sepsis are the main causes of morbidity, the variety of complications is enormous. The literature abounds with reports of morbidity and even death as a result of circumcision.

Operative complications

Haemorrhage and sepsis are the commonest complications and are considered in greater detail below. The nature of circumcision dictates that errors of omission and commission, i.e. too little or too much, is assessing how much foreskin to remove are likely to happen, and one of the commonest complaints is of an unsatisfactory cosmetic result. If insufficient foreskin is removed the cosmetic appearance is such that the penis does not appear to be circumcised; phimosis may still subsequently develop. In a series of consecutive circumcisions in Australia, Leitch found that in 9.5 per cent of patients the operation had to be repeated because of inadequate skin excision at the initial procedure. MacCarthy et al.¹⁷ reported this figure to be 1 per cent in a study from the UK. In a more recent series from Israel, where religious circumcision is widespread, of 60 children referred following potentially inadequate circumcision 42 required recircumcision; the majority of these children were operated on before 4 years of age¹⁸. The rest were treated conservatively and had a satisfactory cosmetic follow-up to 10 years of age. Insufficient excision of the foreskin and inner preputial epithelium may result in wound contraction and cicatrization of the distal foreskin. The fibrotic ring so produced may result in true phimosis, an event observed in 2 per cent of cases in one UK series¹⁵. In severe cases urinary obstruction may ensue¹⁹.

Removal of too much skin from the penile shaft may be caused by pulling it over the glans during operation. After foreskin excision the remaining skin slides back, leaving a denuded shaft. Others suggest that such penile denudation injuries occur as a result of failure to break down the ventral foreskin adhesions to the glans penis completely^12,20. It is therefore essential that, before any incision is made, the inner preputial epithelium be completely free from the glans such that the entire coronal sulcus can be visualized. Penile denudation injuries may occur as a result of sepsis²¹, from diathermy injury²² or from injected substances mistaken for anaesthetic solutions²². The majority of cases can usually be managed conservatively with a satisfactory cosmetic and functional outcome^21,23-25. Such injuries in adults may be managed conservatively if the defect is less than half of the total penile skin. Complete denudation in the adult is managed by split-thickness skin grafting for optimum cosmetic and functional results^21,23,26. Of three cases encountered by Gee and Ansell²⁴, one child with complete denudation had initial treatment by burial of the penis in a tunnel of scrotal skin, no follow-up was available on this patient. Use of a pedicled scotal skin flap has recently been described for the reconstruction of penile shaft skin²⁷.

A rare consequence of excision of excess preputial skin is the so-called 'concealed penis'^23,25,28,29. Kaplan maintains that, although an excess of skin is removed, not enough inner preputial epithelium is excised. The new preputial orifice is therefore, distal to the glans, and as healing and fibrosis occur the penile shaft is forced into the suprapubic fat, with resulting preputial ring at the level of the skin of the mons pubis. Other suggestions about the aetiology of this complication include a tendency of the penis to retract into the mons pubis²⁹ and the possibility that the penile shaft is forced into a subcutaneous position by wound con- traction²⁵. Subsequent fibrosis of the circumcision wound leads to stenosis of the preputial orifice, which then traps the penile shaft subcutaneously^23,29. Treatment of this condition is surgical correction. In the presence of phimosis, Kaplan¹² recommends a circumferential incision at the preputial ring to avoid the need for skin grafts to achieve coverage of the penile shaft. A vertical incision caudal to the circular scar was used by Kon²⁹ to expose the glans and penile shaft which was in a subcutaneous position, tightly adherent to the surrounding tissue. Nearly complete direct skin coverage was achieved, apart from a small ventral defect which was covered by rotation of the scrotal skin. A vertical incision was used also by Byars and Trier²³, and while Shulman et al.²⁵ did not state the incision employed, both required the use of skin grafts for coverage of the penile shaft. More recently, Radhakrishnan and Reyes³⁰ chose to use opposing U shaped incisions and so doing were able to obtain skin coverage with vascularized flaps; they contend that the functional and cosmetic results are superior to those of reconstruction with split-thickness grafts, which may result in comparatively inelastic skin on the penile shaft.

Many other forms of surgical mishap have been reported. Laceration to the penile skin and scrotum resulting in exposure of both testes as reported by Shulman et al.²⁵ was managed by primary suturing. Laceration of the penile shaft with resultant partial amputation has also been described²³. Total ablation of the penis may occur as a result of diathermy injury³¹ and loss of the penis from the use of a rubber band as a tourniquet has been reported³². injury to the glans may result from inadequate separation of preputial adhesions. Glandular injury may be of varying severity and cases of complete surgical amputation of the glans have occurred²⁵. McGowan³³ described a case in which inadvertent placement of scissors into the urethra while attempting a dorsal slit resulted in surgical bivalving of the glans.

Non-Operative Complications

Ruff et al.³⁴ described a case of myocardial injury following immediate postnatal circumcision. This was confirmed by a raised lever of type MB creatine kinase and by left ventricular posterior wall hypokinesis on echocardiography. It was concluded that exposure of the child to cold stress as a result of circumcision resulted in a persistent fetal circulation and hypothermia. The neonate had a satisfactory outcome on medical management. A case of recurrent pneumothorax following circumcision was reported by Auerbach and Scanlon³⁵ in a neonate. The circumcision was complicated by moderate bleeding and the baby's distress was sufficient to produce circumoral cyanosis and persistent tachycardia. It was concluded that crying induced by many dressing changes need to obtain haemostasis, resulted in raised intrapulmonary pressure sufficient to rupture a weak spot and cause pneumothorax. Although the child was managed successfully, this required a further hospital stay of 19 days.

Bleeding

Bleeding remains the commonest complication encountered during and after circumcision. Because varying criteria have been used when recording this complication and because of inadequacies in documentation, the reported incidence ranges from 0·1 to 35 per cent^3,12,24,25. In the majority of cases bleeding is minor and all that is required to achieve haemostasis is gentle pressure on the area. Excessive bleeding may be due to anomalous vessels²⁵ or to the presence of a bleeding disorder²⁴. After 13 000 circumcisions reported in two large series, no patient required blood transfusion for bleeding^24,25. In the event of a bleeding disorder, appropriate clotting factors may have to be administered²⁴. The authors are not aware of any reported cases of exsanguination following circumcision.

In some instances the application of pressure alone is insufficient to control local haemorrhage and other methods of haemostasis must be employed. In the UK the commonest aid to haemostasis is the eletrosurgical diathermy for coagulating vessels. While in the majority of cases the judicious use of this device is safe and effective, the potential for damage exists if it is deployed over zealously. When used in monopolar form, an electrical current flows from the indifferent electrode (plate) to the active electrode (forceps) and the tissue surrounding the forceps is heated, resulting in coagulation. However, this coagulation process may spread proximally in small vessels (a phenomenon commonly observed in everyday practice) and the extent of vessel coagulation may be far greater than anticipated or intended. It is predominantly for this reason that the present authors uso only bipolar diathermy during circumcision. Although there are probably many undocumented cases of minor diathermy burns and sloughing of the affected penile skin, more severe injuries such as glans and major penile skin necrosis have been reported^22,27. At its most severe, the use of diathermy may result in total ablation of the penis. Gearhart and Rock³¹ described four such cases in which damage was so severe that plastic surgical reconstruction was deemed impossible. In all cases the children were managed by gender reassignment and feminizing genitoplasty.

If simple application of pressure is unsuccessful, a type of circumferential bandage may be used to aid haemostasis. This appears to be popular among the practitioners of religious circumcision in Jewish communities. It may cause a degree of urethral obstruction which in severe cases, leads to urinary retention³⁶ and may dispose thus dispose to urinary tract infection³⁶. Horowitz et al³⁸ described such an event in which an 18-day-old infant presented shocked, dehydrated and with a hugely distended abdomen 2 days after circumcision. The tip of the penis, which was still covered by a circular bandage appeared red and necrotic. After the bandage was removed the child voided a large volume of urine and the abdominal distension disappeared. The cause of systemic upset was an Escherichia coli urinary tract infection and subsequent septicaemia. Frand et al.³⁹ described an infant about who the parents were concerned because of a bluish discoloration of the legs 1 day after circumcision; a circular bandage was seen around the penis. After this was removed a large volume of urine was voided and the lower-limb cyanosis rapidly disappeared. It is presumed the extended bladder compressed the iliac veins and so impeded venous return from the legs. Acute urinary retention may be caused by the devices used for circumcision; in one case where the Plastibell device (Abbott Laboratories, Queensborough, UK) was employed, rupture of the bladder was a consequence of acute urinary retention⁴⁰. As well as compressing he urethra the blood flow to the distal penis and glans may also be compromised by the tourniquet action of a circumferential dressing, which in severe cases may result in necrosis of the distal penis and glans⁴¹.

Pharmacological agents may also be used to stop minor troublesome bleeding and in this respect the application of a 1:100 000 adrenaline solution is not uncommon. Although there is a slight danger from systemic absorption, at this low concentration complications are unlikely. If a more concentrated solution is used, however, there is greater systemic absorption with its attendant problems. Mor et al.⁴² described four such cases in which a sponge-soaked solution of 1:1000 adrenaline was sprayed on the the bleeding area of the frenulum after circumcision. The patient developed tachycardia, acrocyanosis and local pallor of the penis. After subcutaneous injection of phentolamine (an alpha-adrenergic receptor antagonist) at the base, shaft and corona of the penis, all penile pallor disappearedand the systemic signs abated.

As an alternative to diathermy, obvious bleeding vessels may be ligated with a fine suture. One of the commonest sites for persistent bleeding is at the frenulum, and in this area it is not uncommon to insert a haemostatic suture. However, because of the close proximity of the underlying urethra, it is easy for such a haemostatic stitch to be placed in the urethra itself. The result is development of a urinary fistula^44,45. Avoidance of this complication depends on meticulous technique when suturing around the frenulum and in taking superficial tissue only in the stitch.

Sepsis

Infection occurs after circumcision in up to 10 per cent of patients^3,15. In the majority of cases this is usually mild and manifested by local inflammatory changes, but occasionally there is ulceration and suppuration. Most infections are of little consequence and settle with local treatment, Occasionally, however, sepsis may have a more alarming consequences and may even cause death^46-48.

The perineal skin is heavily colonized by both normal skin saprophytes and by bowel flora, and it is surprising that significant septic complications do not occur more frequently. Woodside^49,50 reported a case of necrotizing fasciitis of the perineum after Plastibell circumcision of a neonate. The child was sepicaemic and there was also bacterial contamination of the cerebrospinal fluid. Extensive surgical debridement of the perineum with multiple fasciotomy was necessary and parenteral antibiotics were administered. Staphylococcus aureus, S. epidermidis, diphtheroids, alpha-haemolytic streptococcus and Clostridium perfringens were cultured from the excised perineal and penile skin. The child survived and had a satisfactory cosmetic and functional result some years later.

Gangrene of the penis has also been reported following circumcision⁵¹ and the scope of microbial contamination of the perineal skin can be further appreciated from the report of Sussman et al.⁵² of two cases of Fournier's syndrome of the perineum and scrotum following the operation. Both patients survived after adequate antibiotic management, although extensive surgical debridement was necessary in one child. Annunziato and Goldblum⁵³ presented three cases of staphylococcal scalded skin syndrome all of which occurred after circumcision and similarly cases of impetigo have been documented⁵⁴. In 1935, Gosden⁴⁷ reported a series of 23 boys from Cyprus circumcised by an unqualified practitioner in 13 of whom tetanus developed; five died. Diphtheria may develop in a circumcision wound⁵⁵ and infection with Mycobacterium tuberculosis has been documented¹⁸. The circumcision site has been observed to be the portal of entry in many other cases of sepsis, and Cleary and Kohl⁵⁶ reported a fatal case of septicaemia with group B beta-haemolytic streptococcus in a 6-week-old child. Kirkpatrick and Eitzman⁵⁷ highlighted the significance and occurrence of septicaemia following the circumcision with the Plastibell device. Metastatic infection from the circumcision site has been reported as a cause of neonatal meningitis in five children^58,59, three of whom were treated successfully with no long-term sequallae. However, one child subsequently developed cerebral palsy and one died. Osteomyelitis of the femur²⁵ and a fatal case of staphylococcal bronchopneumonia⁶⁰ have also been reported following circumcision. Recently, Crowley and Kesner⁴⁸ reported a 9 per cent mortality rate in a series of 45 consecutive admissions for septic complications following ritual circumcision in young adults. Septic embolization and polyarthritis were among the documented complications of septic circumcision in this series. Despite its apparently low incidence, sepsis following circumcision has the potential to cause significant morbidity and in some cases death.

Fistula

Urethrocutaneous fistula following circumcision may occur for a variety of reasons but, fortunately the reported incidence of this complication is low. Perhaps the commonest cause is a poorly placed suture at the frenulum in an attempt to obtain haemostasis^44,45. This results in strangulation and necrosis of part of the urethral wall, with resultant subglandular fistulation not dissimilar to glandular hypospadias. Fistulation may also occur as a result of sepsis⁶¹ or unrecognized rare penile anomaly, such as megalourethra⁶². However, many other fistulas arise from using the Plastibell device or Gomco clamp^12,44. Although the mechanism of injury is not clear it is probable that urethral injury results from crushing by the device. Most of these fistulas open on to the dorsum of the penis but they may open to the ventral surface), an anatomical arrangement not dissimilar to that seen in epispadias. Such a case was reported following surgical bivalving of the glans³³. Although there are many approaches to the management of such a urethral fistula⁶³, appraisal of the techniques employed is beyond the scope of this review.

Meatal stenosis

Meatal stenosis is generally a direct consequence of circumcision that is seldom encountered in uncircumcised men; meatal calibre is know to be greater in uncircumcised individuals. The incidence of meatal ulceration following circumcision is from 8 to 20 per cent^14,46,64. The aetiology is thought to be irritation of the external urethral meatus by ammoniacal substances present in wet sodden nappies. Such irritation is unlikely in the presence of a normal prepuce, which serves to protect the glans from these irritant substances⁴⁶. In a prospective study of 140 consecutive neonatal circumcisions, Mackenzie⁶⁴ found a 20 per cent incidence of meatal ulceration within the first 2-3 weeks after circumcision. It is thought that meatal ulceration after circumcision is the initiating event in a vicious cycle of stenosis and ulceration, followed by more stenosis⁶⁴. Meatal stenosis following circumcision has been advanced as a cause of recurrent pyelonephitis and obstructive uropathy, for which meatotomy is curative^64,65.

Miscellaneous complications

The glandular epithelium may be denuded by a less than gentle technique when separating the preputial adhesions and this may be exacerbated if the glans is held firmly by a gauze swab. The glans may then be more prone to local sepsis with the resulting formation of a scab. Fortunately, most settle spontaneously with attention to hygiene. A skin bridge may develop between the glans and the penile shaft. This may tether the erect penis so producing pain and deformity. The aetiology of this condition remains unknown, although injury to the glans at the time of circumcision or incomplete separation of the inner preputial skin have been advanced as possible factors^12,66. Inclusion cysts following circumcision have been described. That reported by Shulman et al.²⁵ was found histologically to be an epidermal cyst. It is possible that cysts also arise as a result of peroperative implantation of smegma. The use of silica talc on surgical gloves has been associated with the formation of granulomas and such a lesion was described in a circumcision wound 15 years after the original surgical procedure⁶⁷. Although penile lymphoedema following circumcision has been reported, there is a paucity of information regarding the aetiology and management of such a problem and accounts in the literature are anecdotal^12,25. The degree of penile oedema may be greater with the Plastibell device⁶⁸. Impotence has been observed following circumcision⁶⁹. Palmer and Link⁷⁰ described two cases both of which were associated with injection of 1 per cent lignocaine directly into the corpora after application of a rubber band at the base of the penis to act as a tourniquet. As 10 - 15 ml anaesthetic was used this perhaps resulted in a 50 per cent mixture of lignocaine and blood in direct contact with the vascular endothelium of the penis. It is postulated that this high concentration irreversibly damaged the endothelium of the corpora cavernosum with resulting impotence. Impotence followed partial amputation of the penis at circumcision, despite plastic reconstruction, in a case reported by Hanash⁷¹. Hypospadias remains a contraindication to circumcision, as surgical reconstruction may require the use of all available penile skin. However circumcision in unrecognized hypospadias was performed in six children in one series and, in the same, series, although hypospadias was recognized circumcision was performed anyway.

Carcinoma

Of 1103 patients with carcinoma of the penis reviewed by Wolbarst⁷², none had been circumcised. It was therefore concluded that squamous carcinoma of the penis only ever occurs in uncircumcised men and that the circumcised state is protective against its development. However, squamous cell carcinoma of the uncircumcised penis has since been reported^73,74 and so the original contention of a protective effect is not convincing. Indeed, when one compares the incidence of carcinoma of the penis in the USA (0-2·1 per 100 000) where most of the men have been circumcised at birth, with that in Denmark (1·1 per 100 000) and Japan (0·3 per 100 000), where circumcision is rarely performed, the original contention seems doubtful^5,75. It is likely that several factors other than circumcision are implicated in the genesis of carcinoma of thepenis.

Carcinoma of the penis following circumcision appears to have a different natural history from cancer in uncircumcised men. Whereas penile cancer in the uncircumcised tends to arise on the glans or prepuce, after circumcision the tumour is likely to develop in the surgical scar. Such tumours occur mostly on the penile shaft and tend to spread locally with distant metastasis as an infrequent or late occurrence⁷⁶. Surgical excision is the treatment of choice, as neither radiotherapy nor chemotherapy appears to be effective⁷⁷. Nearly all of these cases have been reported from Saudi Arabia, where there existed a radical form of circumcision that was practised by the tribes in the southern regions. The circumcision included excision of the skin in the suprapubic region with a longitudinal incision extending between the anterior iliac spines⁷⁸. This practice is now prohibited.

Psychological considerations

According to Freudian theory, by the fourth or fifth year of like the genital concentration of all sexual excitement is achieved and the boy's interest in the genitals attains a dominant significance, the phallic stage. Anna Freud⁷⁹ wrote:

`any surgical interference with the child's body may serve as a focal point for the activation, reactivation, grouping, and rationalisation of ideas of being attacked, overwhelmed and (or) castrated.'

For a child at the phallic stage this fear that something might happen to his prized organ is called 'castration anxiety'. According to the psychological literature. operations performed on the penis, such as circumcision, may arouse such castration fears. In a study on circumcision, and the problems of bisexuality, Nunberg⁸⁰ highlighted the point that injury to the penis may intimidate the child and impair his development to full virility. However in the same paper he also proposed that circumcision may stimulate the masculine striving of the child by encouraging identification with the father. After a detailed study of 12 boys aged 4-7 years undergoing circumcision in Turkey, Cansever⁸¹ found that following surgery there was a significant fall in intelligence quotient and that body image showed a tendency to contract. She concluded that:

`circumcision is perceived by the child as an aggressive attack upon his body, which, damaged, mutilated, and in some cases destroyed him. The feeling of I am now castrated seems to prevail in the psychic world of the child. As a result he feels inadequate, helplessand functions less efficiently.'

Such psychological sequalae are not confined to the young child. Circumcision performed in the neonatal period is associated with marked behavioural changes^82,83 that may last up to 24 h; supporting this observation is the finding of a rise in both serum cortisol and cortisone levels after neonatal circumcision⁸⁴. Circumcision in the neonate has been noted to increase both respiration and heart rate. and is associated with a significant fall in transcutaneous oxygen tension⁸⁵. Allied to this is as change in sleep pattern with prolonged non-rapid eye movement sleep. This change has been interpreted as 'being consistent with a theory of conservation - withdrawal to stressful stimulation'⁸⁶.

Circumcision can cause dysmorphophobia, and Walter and Streimer⁸⁷ reported a case of genital self-mutilation in a non-psychotic patient who attempted to reconstruct the foreskin himself. Two groups of men seeking restoration of the foreskin have been identified. First is a group of Jewish men who sought to disguise their religious identity during times of political crisis⁸⁸. Second is a group of homosexuals whose circumcised status is associated with unwanted masculinity and anger over having no choice over their circumcision⁸⁹. Genital mutilation by attempted circumcision has been reported after paternal death in two patients, both of whom exhibited features of acute psychosis⁹⁰. Schizophrenia following elective circumcision has also been reported.⁹¹.

In societies in which circumcision is intricately linked to tradition and culture, the uncircumcised individual is likely to be an outcast. This prejudice may be great enough for uncircumcised men not only to be ostracized by their peers but even to be attacked and beaten for their lack of conformity. Such beatings in men refusing to be circumcised have occurred in the Xhosa tribe of South Africa and, in one instance, the attack was violent enough to result in the development of 'crush syndrome'⁴⁸.

Sexual Complications

Morgan⁹² asserts that:

`penetration in the circumcised man has been compared to thrusting the foot into a sock held open at the top while, on the other hand, in the intact counterpart it has been likened to slipping the foot into a sock that has been previously rolled up'.

He also suggests that 'coitus without a foreskin is comparable to viewing a Renoir while colour blind'⁹³. While it is impossible to reach an objective conclusion in his matter, some have risen to the challenge and have attempted to answer the question. Although Harnes⁹⁵ concluded the question remains unanswered at the end of his study, the description of his methodology and findings make for compulsive reading.

Conclusion

Many medical practitioners regard circumcision as a relatively minor procedure and, as such, it is likely to be delegated to a junior surgeon. It has been observed that the complication rate is directly related to operator inexperience³. Delegation to a junior colleague should occur only after the surgeon in training is fully instructed in the operative procedure. Heightened awareness of the scope and potential for complications will of itselfresult in a reduced complication rate.

Despite the warnings of the British Medical Association that circumcision should be performed only for medical reasons, controversy still exists over whether too many procedures are being carried out. It is hoped that greater awareness of the incidence and scope of associated complications will encourage a more carefully considered decision on whether or not to circumcise.

Acknowledgements

The authors thank the library staff at the Frank Rifkin Postgraduate Centre, Hope Hospital, for help in obtaining many references and Mrs. Anne Bradley for assistance in preparing the manuscript.