Prevention and management of BK-virus associated haemorrhagic cystitis in children following haematopoietic stem cell transplantation – a systematic review and evidence-based guidance for clinical management


Dr C. Harkensee, Children’s Bone Marrow Transplant Unit, Newcastle General Hospital, Westgate Road, Newcastle Upon Tyne NE4 6BE, UK. E-mail:


Haemorrhagic cystitis (HC) is a common and, in its severe form, potentially life-threatening complication of Haematopoietic stem cell transplantation (HSCT) in children. Recent data indicate an important role of BK virus reactivation during the time of maximal post-transplant immune suppression in the pathogenesis of late-onset HC. Treatment of HC is mainly symptomatic and often frustrating. To give clinicians guidance on prevention and treatment options and their backing by scientific evidence, we have systematically assessed the available literature and devised evidence-based guidelines. Our comprehensive review demonstrates that evidence for the most commonly used interventions (such as cidofovir, oestrogen, hyperbaric oxygen, bladder instillation with formalin, alum salts or prostaglandin) is very limited. Some of these interventions also carry significant risks. Higher level evidence exists only for 2-mercaptoethane sodium (MESNA) and hyperhydration as a preventative intervention, and for systemic recombinant Factor VII as a treatment to stop acute haemorrhage. Further high-quality studies are required to establish effective and safe prevention and treatment options for HC.

Haemorrhagic cystitis (HC) is characterised by haemorrhagic inflammation of the bladder mucosa, leading to painful micturation with haematuria. HC is commonly associated with immunocompromise after chemotherapy or irradiation, or following haematopoietic stem cell transplantation (HSCT). Clinically, HC can be mild and brief (Grade I), ranging to severe, prolonged and life-threatening (Grade IV) (Table I) (Bedi et al, 1995; Iwamoto et al, 2002).

Table I.   Clinical grading of HC (Droller et al, 1982).
GradeClinical features
IMicroscopic haematuria
IIMacroscopic haematuria
IIIMacroscopic haematuria with blood clots
IVMacroscopic haematuria, blood clots, urinary obstruction, renal impairment, bladder impairment

There are two major forms of HC post-HSCT: Early-onset (within days after transplantation), pre-engraftment disease, usually mild, and associated with the conditioning regimen (chemotherapy, irradiation); and a late-onset, post-engraftment disease, associated with the reactivation of urotropic viruses (BK virus [BKV], Adenovirus, cytomegalovirus [CMV]) (Leung et al, 2002).

The incidence and severity of pre-engraftment HC has decreased due to less toxic conditioning regimens, uroprotective antitoxic agents [e.g. 2-mercaptoethane sodium (MESNA)] (Hows et al, 1984) or hyperhydration/forced diuresis regimens (Shepherd et al, 1991), but post-engraftment HC remains a major challenge with respect to pathogenesis, diagnosis and management.

Post-engraftment HC could be regarded as ‘post-engraftment disease’, with onset usually within a month of neutrophil engraftment, followed by variable disease severity and duration (1 week to 4 months) (McCarville et al, 2000). These features suggest multiple contributing risk factors, such as pretransplant viral status, conditioning and immunosuppressive regimens, degree of cellular immunosuppression, Graft-versus-Host Disease (GvHD) and host/donor genetics. Leung et al (2005a) have suggested a 3-phase model for post-engraftment HC: uroepithelial insult by chemotherapy and radiation providing a permissive environment for virus replication (phase 1), which is reactivated in the face of immunosuppression (phase 2). Upon engraftment, infected uroepithelial cells are attacked by donor lymphoid cells, leading to tissue destruction (phase 3). It has been previously argued that post-engraftment HC could represent uroepithelial GvHD (Ost et al, 1987). It is likely that human leucocyte antigen (HLA) and immune response gene polymorphisms play a role in viral immune responses, as shown in BKV nephropathy (Ellis et al, 2004; Bohl et al, 2005).

BK virus

BK virus, a member of the Polyomaviridae family first described in 1971, was isolated in cell culture from the urine of an asymptomatic immunosuppressed patient (Gardner et al, 1971). Primary infection with BKV usually occurs in childhood and is generally asymptomatic. Thereafter the virus lies latent in the host. BKV is urotheliotropic, affecting epithelia of renal calyces, renal pelvis, ureter and urinary bladder. The widespread frequency of BKV in children suggests common routes of transmission such as respiratory or faecal spread.

Diagnosis and risk factors for BKV infection in HSCT

It is now generally believed that BKV infection during HSCT is due to reactivation of latent virus, although new or re-infection has also been postulated (Bogdanovic et al, 1998; Dolei et al, 2000; Bohl et al, 2005; Leung et al, 2005b).

Early studies associating BKV with HC were inconclusive (Arthur et al, 1986; Cottler-Fox et al, 1989; Chan et al, 1994; Rabenau et al, 2002). Urinary BKV load can be quantified by polymerase chain reaction; urinary BK viral load peaks correlate with subsequent development of HC (Azzi et al, 1999; Leung et al, 2001, 2005b; Bogdanovic et al, 2004). Blood viral titres >104 copies/ml preceded onset of HC by a median of 10 d (Erard et al, 2004, 2005). A prospective observational study (Cesaro et al, 2007) found a urine BKV load of >9 × 106 copies/ml and blood BKV load >1 × 103 copies/ml predictive of HC in children, with a higher sensitivity for urine monitoring. Viruria preceded onset of clinical HC by a median of 18 d, viraemia by 17 d.

The incidence of BKV-associated HC varies across different transplant populations, ranging from 3·6% to 20%, using different definitions for HC. A number of prospective and retrospective case series have investigated risk factors for the development of post-engraftment HC, including myeloablative conditioning (Giraud et al, 2006), unrelated donor transplants (Trotman et al, 1999; Chakrabarti et al, 2003; Bogdanovic et al, 2006), and Adenovirus and CMV infection (Akiyama et al, 2001; Asano et al, 2003; Yamamoto et al, 2003; Tomonari et al, 2006). Demographic risk features include male sex (Asano et al, 2003; Hale et al, 2003) and age >10 years (Kondo et al, 1998; Seber et al, 1999; Cesaro et al, 2003; Cheuk et al, 2007). GvHD is a consistent risk factor in paediatric (Russell et al, 1994; Hale et al, 2003; Cheuk et al, 2007) mixed (Seber et al, 1999; El-Zimaity et al, 2004; Hassan et al, 2007) and adult (Chakrabarti et al, 2003) study cohorts. Busulphan (Kondo et al, 1998; Trotman et al, 1999) and cyclophosphamide conditioning (Brugieres et al, 1989; Cheuk et al, 2007), important risk factors for pre-engraftment HC, also seem to increase risk of post-engraftment HC. Immunosuppressive therapies, including T-cell depletion, antithymocyte globulin, methotrexate, cyclosporin and tacrolimus all lead to a higher incidence of HC (Childs et al, 1998; Trotman et al, 1999; Hale et al, 2003; Brennan et al, 2005; Agha & Brennan, 2006; Wu et al, 2006).

Aims, objectives and methodology

This systematic review aimed to identify the best available published evidence with regards to the prevention and treatment of BKV-associated post-engraftment HC in children.

Details of the search protocol are available in Appendix SI. Briefly, we conducted unlimited searches, using different homonymous search terms, in three databases (Medline/Pubmed from 1950/1966 onwards, Embase, Cochrane databases of systematic reviews and clinical trials). The results were categorised into study types, and quality assessed and graded (Harbour & Miller, 2001) using appropriate forms on the SIGN website. Searches and assessments were performed independently by two authors each, separately for medical and surgical interventions. Results of selection and assessment were discussed within teams and inclusion/exclusion of studies agreed.

Proceeding through searches and assessments of the literature, it became obvious that our initial aim of presenting a firm evidence-based guideline based on a reasonable number of high-quality studies became unachievable because of the apparent lack of such studies. Therefore, this report presents a comprehensive, systematic review and grading of the available literature with an informative character about management options, rather than conclusive recommendations based on hard evidence.


Studies included

Evidence for prophylaxis and management of BKV-associated post-engraftment HC in children is limited. We included six randomised controlled trials (RCT) (Hows et al, 1984; Shepherd et al, 1991; Praveen et al, 1992; Vose et al, 1993; Bedi et al, 1995; Pihusch et al, 2005). Details are summarised in Table SI. Although some represent high quality evidence (grading 1+), for the purpose of recommendation evidence had to be downscaled because none would fit fully our initial inclusion criteria, (child – post-engraftment HC after HSCT – BKV association). These studies contributed to a recommendation grade higher than ‘D’ (see Appendix I for a definition of evidence and recommendation grades) for only two interventions: the equal effectiveness of MESNA and hyperhydration/forced diuresis, and short-term control of haematuria using recombinant activated factor VII (rFVIIa). The remainder of evidence consisted of non-randomised and/or non-controlled, small clinical trials and prospective/retrospective case series; by SIGN criteria evidence level 3, hence resulting in a recommendation grade ‘D’. The identified evidence was grouped by type of intervention, focusing on interventions commonly applied and for which high level evidence was identified. Results are summarised in Table II. Less commonly used interventions were not described in detail, but a summary, recommendation and references of these are included in Table II.

Table II.   Evidence based guidance: evidence level and grade of recommendation for preventative and medical therapeutic interventions.
InterventionEffectAdvantageDisadvantageGrade of RecommendationLevel of evidence, Reference
Preventative interventions
 CiprofloxacinReduction of viruriaLow toxicityProbably mild efficacyD3 (Leung et al, 2005a)
 MESNA/hyperhydrationEqually effective in preventing HCLow toxicityHypersensitivity (rare)B1+ (Shepherd et al, 1991; Vose et al, 1993; Bedi et al, 1995)
1 (Hows et al, 1984)
3 (Ballen et al, 1999)
 Continuous bladder irrigationPrevention of HCLow toxicityRequires elective catheterisationD3 (Hadjibabaie et al, 2008)
 Sorbitol irrigationPrevention of HCLow toxicityRequires elective catheterisationD3 (Rosenzweig et al, 1994)
Therapeutic interventions – systemic
 Recombinant FVIICessation of acute bleedingEffectiveHigh cost, potential for thrombembolisms, short effect durationB1+ (Pihusch et al, 2005)
3 (Blatt et al, 2001; Karimi et al, 2004; Ashrani et al, 2006)
 Recombinant FXIIICessation of acute bleedingProbably effectiveHigh cost, potential for thrombembolismsD3 (Sakuma et al, 1994; Demesmay et al, 2002)
 Cidofovir (systemic)Causative against BKVEffective in renal transplant BK nephropathyNephrotoxicD3 (Gorczynska et al, 2005)
 Cidofovir, low dose(systemic)Causative against BKVEffective potentially less nephrotoxicProbably less effective than full dose cidofovirD3 (Savona et al, 2007)
 VidarabineCausative against BKVProbably effectiveNephrotoxic, potentially less effective than cidofovirD3 (Kawakami et al, 1997; Seabra et al, 2000; Vianelli et al, 2000)
 OestrogenPromotes wound healingProbably mildly effectiveHepatotoxicD3 (Liu et al, 1990; Ordemann et al, 2000; Heath et al, 2006; Cheuk et al, 2007)
 Hyperbaric oxygen therapyPromotes wound healingProbably effective, safeAvailability, expensiveD3 (Hattori et al, 2001; Cesaro et al, 2003)
 Recombinant KGFPromotes wound healingPotentially effective, safeExpensiveD3 (Czibere et al, 2007)
Therapeutic interventions – local
 Intravesical prostaglandinCessation of acute bleedingProbably effective, no systemic effects reportedBladder spasmsD1 (Praveen et al, 1992)
3 (Trigg et al, 1990; Levine & Jarrard, 1993; Ippoliti et al, 1995; Laszlo et al, 1995; Cesaro et al, 2003)
 Cidofovir instillationCausative against BKVPotentially less nephrotoxic D3 (Bridges et al, 2006; Walden et al, 2007)
 Sodium hyaluronate instillationCessation of bleedingNon-toxic D3 (Miodosky et al, 2006)
 Alum instillationCessation of bleedingProbably effectiveNephrotoxic, systemic toxicityD1 (Praveen et al, 1992)
3 (Goswami et al, 1993; Efros et al, 1994)
 Formalin instillationCessation of bleedingProbably effectiveHighly toxic, high rate of complications, systemic toxicityD3 (Shrom et al, 1976; Redman & Kletzel, 1994; Cheuk et al, 2007)
 Silver nitrate InstillationCessation of bleeding, antisepticProbably effectiveSilver salt deposition, systemic toxicityD3 (Kumar et al, 1976)
 Growth factor instillationCessation of bleedingProbably effective, non-toxicExpensiveD3 (Vela-Ojeda et al, 1999; Dorticos et al, 2003; Czibere et al, 2007)
Surgical interventions
 Fibrin glueCessation of bleedingProbably effective non-toxicRequires cystoscopy in general anaesthesiaD3 (Purves et al, 2005)
 Selective embolisationCessation of bleedingEffectiveInterventional, risk of vascular complicationsD3 (Gine et al, 2003)
 Intravesical hydrostatic pressure balloonCessation of bleedingEffectiveRequires general anaesthesia, risk of traumatic complicationsD3 (Helmstein, 1972; England et al, 1973; Hammonds et al, 1974; Hansen et al, 1976)
 Cystostomy/cystectomyCessation of bleedingEffectivePermanent changes to anatomy and functionD3 (Garderet et al, 2000)


MESNA/hyperhydration.  Hyperhydration with forced diuresis has been studied in context of pre-engraftment HC caused by toxic metabolites of cyclophosphamide or ifosfamide. The uroprotective effect of MESNA as an uroprotective antitoxic agent has been studied as part of high-quality chemotherapy drug trials, but also in efficacy, tolerability and safety compared with hyperhydration (Shepherd et al, 1991; Ballen et al, 1999) or prophylactic bladder irrigation(Vose et al, 1993). Results are equivocal, with only one trial reporting an advantage of MESNA over forced diuresis/hyperhydration (Hows et al, 1984). MESNA and hyperhydration appear to be equally effective in preventing HC, although studies did not distinguish between early and late onset HC, therefore the protective impact on post-engraftment BKV-associated HC cannot clearly be determined.

Continuous bladder irrigation.  The prophylactic insertion of a urinary catheter and continuous irrigation of the bladder with normal saline during the administration of high dose chemotherapy or conditioning regimens was common practice before the introduction of MESNA, with controversial results (Atkinson et al, 1991; Turkeri et al, 1995). A very recent, non-randomised clinical trial in the HSCT setting (Hadjibabaie et al, 2008) consisted of a prospective group of HSCT patients (n = 40) receiving continuous bladder irrgation with normal saline for 36 h during the conditioning regimen in addition to MESNA, hyperhydration and urine alkalinization, and a historical, retrospective control group (n = 40) of similar characteristics and management, but no preventative bladder irrigation. A significant reduction in incidence and duration of pre-engraftment HC was observed. In addition, also the incidence of late-onset HC (here defined as occurring after day 30 post-HSCT was significantly decreased in the bladder irrigation group (7·7%, compared to 45% in the control group, P = 0·009). No data on BKV status were recorded.

Ciprofloxacin.  The flouroquinolone antibiotics have a known in vitro activity against BKV (Ferrazzi et al, 1988; Portolani et al, 1988). A single non-randomised clinical trial (Leung et al, 2005a) investigated the effect of prophylactic ciprofloxacin (500 mg twice daily oral/200 mg twice daily iv) on BK viruria and incidence and severity of post-engraftment HC in adult patients after HSCT. There was a significantly lower BK viruria peak in the group receiving ciprofloxacin, but no demonstrable reduction in HC incidence or severity.

Medical treatment – systemic interventions

Cidofovir. has the highest specifity against BKV (Andrei et al, 1997). Significant nephrotoxicity limits routine use following HSCT where patients are usually concomitantly exposed to other nephrotoxic drugs. In the context of BKV infection, most experience stems from the renal transplantation setting (Vats et al, 2003). No RCT or prospective clinical trial so far has evaluated its role in children with post-engraftment BKV-associated HC, although it is widely used for this indication.

A prospective case series (Gorczynska et al, 2005) observed a cohort of 102 paediatric HSCT patients for virus-associated HC. Cidofovir was given to 22 patients at a dose of 5 mg/kg initially twice weekly, then fortnightly, leading to complete erradication from blood, but not urine. No nephrotoxicity was observed.

Treatment with low dose cidofovir (1 mg/kg) was retrospectively reviewed (Savona et al, 2007) in 19 adult patients with BKV post-engraftment HC. 84% of patients experienced clinical improvement, but less than half had microbial resolution. Although rate of response seems to be lower and time to response longer, compared to full-dose studies, low dose cidofovir may represent an alternative in patients with renal impairment.

A number of case reports describe successful cidofovir treatment of concomitant BKV HC with CMV (Held et al, 2000) and HSV(Andrei et al, 2007). More recently, Walden et al (2007) and Bridges et al (2006) reported cases of intravesical administration of cidofovir as an alternative to systemic therapy, leading to clinical improvement and reduction of viruria.

Oestrogen.  Oestrogen has anti-inflammatory properties and cases of successful treatment with oestrogen were first reported for chemo- or radiotherapy-associated cystitis (Liu et al, 1990), (Miller et al, 1994). A small number of case series describe d the treatment of post-engraftment HC. IV. administration of oestrogen at doses of 25–100 mg/d, followed by oral maintenance therapy with 5–10 mg/d (Heath et al, 2006) seemed to be more effective in reducing symptoms than oral therapy with 6–12 mg/d alone (Cheuk et al, 2007), (Ordemann et al, 2000). Side effects were rare, although cases of hepatosis requiring treatment withdrawal have been reported (Ordemann et al, 2000).

Hyperbaric oxygen.  Hyperbaric oxygen therapy (HOT), through an increased oxygen pressure in tissues and stimulation of neovascularisation, has a healing effect on necrotic or injured tissues. This effect has been exploited extensively for cystitis, especially radiotherapy-induced HC, but a recent Cochrane review (Denton et al, 2002) failed to identify high-level evidence for its efficacy. The largest series of children treated with HOT for post-engraftment HC stems from a multi-centre retrospective case series (Cesaro et al, 2003). HOT was administered for 14 episodes of HC, all patients had grade II-III disease, five had previous prostaglandin E2 (PGE2) treatment. The treatment consisted of daily sessions of HOT, with remission occurring in 78·5% of patients at a median of 17 d (4–58 d). HOT was significantly more successful in resolving HC than PGE2 (see below). Further paediatric cases of successful treatment of HC post-transplant/chemotherapy have been reported (Hattori et al, 2001), (Furness et al, 1999) Side effects are rare, but have been documented (Al-Seraihi & Ayas, 2007). HOT is a non-toxic treatment option, limited by the availability of decompression chambers and suitability for sick patients.

Clotting factors. Recombinant activated Factor VII has a haemostatic effect leading to formation of thrombin and a haemostatic plug. rFVIIa was investigated in a randomised, placebo-controlled clinical trial (Pihusch et al, 2005). This study enrolled 100 patients aged >12 years with bleeding complications between days 2–180 post-HSCT, 26 of those had HC. rFVIIa was given in three different doses (40, 80 and 160 μg/kg) as seven single administrations over a 36-h period, and compared with placebo. Overall, a reduction of the bleeding score at endpoint (38 h after first administration) was observed for 80 μg/kg, but not for 160 μg/kg. Six thrombembolic events, including two deaths, were attributed to the study medication.

A different dose regimen was used in a prospective case series (Ashrani et al, 2006) of patients with HC after high-dose chemotherapy. Seven adult patients received doses of initially 80 μg/kg, followed by two further administrations of 120 μg/kg at 3-h intervals if bleeding persisted. Four patients had complete, and a further two partial responses which were all short lasting (bleeding recurred to baseline within hours). Two further small case series reported doses of 100 and 400 μg/kg (Karimi et al, 2004) and 90 and 270 μg/kg (Blatt et al, 2001) to be effective in HC.

As a standard rFVIIa dose of 90 μg/kg costs around £4000 in the UK, treatments according to the above study protocols would cost between £12 000 (for three doses) and £28 000 [for seven doses, source: British National Formulary (BNF) 2006].

Factor XIII is important for wound healing, and its use has been described in two small case series. Sakuma et al (1994) reported four children with severe HC post-HSCT or high dose chemotherapy, of which two had low Factor XIII levels. Patients received doses between 20 and 230 IU/kg and all experienced complete remission of haemorrhage. In the second series, Demesmay et al (2002) described four adult patients with post-engraftment HC (one with proven BK reactivation, and one with a low FXIII level) who received a standard dose of 50 IU/kg of FXIII. Three patients had an initial remission.

Keratinocyte growth factor.  Keratinocyte growth factor (KGF) was demonstrated to have the capacity of preventing HC after cyclophosphamide administration in an animal model (Ulich et al, 1997). Most recently, Czibere et al (2007) reported a single case of a 24-year-old patient with severe, BKV-associated HC following HSCT who responded to parenteral recombinant KGF. Doses of 60 μg/kg (price for a single dose: approximately £544, BNF 2006) were given as five courses of three injections/week for full remission.

Bladder instillations and irrigations

Intravesical prostaglandin instillation.  Prostaglandin is thought to encourage platelet aggregation and cause contraction of vascular smooth muscle on bladder mucosa. A small randomized controlled trial (Praveen et al, 1992) of 19 adult patients with HC caused by bladder carcinoma or radiation cystitis compared the bladder instillation of 1 mg daily of 15(s) 15 Me PGF2 alpha for 5 d (10 patients) with the continuous irrigation with 1% alum over 72 h (nine patients). In each arm, six patients reached full remission, although in the alum arm more patients had recurrence of bleeding. Vesical spasm as a side effect occurred at similar frequency and intensity in both arms of the trial and responded to antispasmodics.

A number of case series, mainly in adult patients with HC after HSCT, studied different dosages and ways of administration, with very similar outcomes. Instillation of either 0·75 mg/d of PGE2 (Laszlo et al, 1995), 500 μg/d (Nakaseko et al, 1995) or 375 μg/d of PGE1 (Trigg et al, 1990), or 0·4–1·0 mg of intravesical carboprost tromethamine four times/day lead to resolution of symptoms in 80–100% of patients.

In children, a review of 1217 cases of HSCT between 1983 and 1999 (Cesaro et al, 2003) identified 44 cases of HC, of which 39 were post-engraftment. The efficacy of PGE2 was compared with hyperbaric oxygen therapy (HOT), finding that HOT was more efficient in resolving haematuria than PGE2 (P = 0·002).

A small prospective dose escalation trial(Ippoliti et al, 1995) for continuous irrigation with carboprost on 24 adult patients in a HSCT setting compared different peak doses, which were escalated from 0·2 mg/dl over 1 h up to 1·0 mg/dl over 1 h. Doses of 0·8 mg/dl per hour appeared to be a cut-off for a minimum effective dose.

Intravesical alum salt instillation.  Alum salts, when administered as a 0·5–1% solution by continuous bladder irrigation, precipitate surface proteins, decrease capillary permeability, and lead subsequently to a reduction in inflammation, oedema and exsudates (Arrizabalaga et al, 1987).

Indicated by a large numbers of case reports, irrigation of the bladder with aluminous salts is a very commonly used measure in children and adults with intractable bladder haemorrhage. Nevertheless, only a single small RCT (Praveen et al, 1992) (see above) has studied its effects in a systematic way in the context of HC. This study found intravesical prostaglandin and continuous bladder irrigation with alum to be similarly effective in inducing remission of haematuria, although alum treatment had a higher incidence of relapse. There is a number of case reports reporting severe aluminium toxicity presenting as encephalopathy (Kanwar et al, 1996; Phelps et al, 1999; Nakamura et al, 2000), even leading to death (Seear et al, 1990; Shoskes et al, 1992). These events mostly occurred in the context of renal impairment, making alum salts a drug to be used with caution. Prothrombin time and Aluminium levels are useful for monitoring toxicity (Goswami et al, 1993). We identified only one retrospective case series in patients with HC following HSCT (Efros et al, 1994). Only one of the six patients with HC responded to alum.

Intravesical formalin instillation.  On the mucosal bladder, formalin precipitates proteins, and occludes and fixates teleangiectatic tissue. Formalin instillations of the bladder were introduced in the late 1960’s for the treatment of radiotherapy cystitis (Brown, 1969) using high concentrations (10% formalin solution), achieving success rates of >90%.

Formalin is instilled under general or spinal anaesthesia. The concentration of formalin remains the most important variable, as the complication rate increases with increasing concentration, reaching up to 75% for a 10% solution, while complications are rarely observed using 1–2% solutions (Fair, 1974). Complications include local effects, such as irreversible fibrosis of the urinary tract, bladder perforation, renal failure, and systemic toxicity on the myocardium and central nervous system.

Three small series using local formalin treatment included children. One of those used irrigation of the bladder with 4% formalin for 10–20 min for HC in the HSCT setting (Cheuk et al, 2007); the other two (Shrom et al, 1976) (Redman & Kletzel, 1994) used 4% formalin instillation for postchemotherapy HC. All achieved high remission rates (4/5, 3/3 and 5/6 patients, respectively). Two patients within these series had formalin-related complications (pyelonephritis, hydroureteronephrosis).

Substantial evidence regarding the significant risk of toxic side effects comes from studies on radiotherapy-related HC in adult populations. In a drug escalation study (Dewan et al, 1993), administering 1%, 2% and 4% formalin in three different arms, no significant difference was seen in the response rate. The higher concentrations of formalin, however, lead to an increased risk of major complications, which were 31% overall.

Intravesical instillation of growth factors.  A number of cases have been reported for intravesical irrigation with recombinant growth factors for the treatment of intractable HC after HSCT. Vela-Ojeda et al (1999) reported a series of six adult patients with late-onset HC, of which five patients had complete responses within 36 h after daily instillation of 400 μg GM-CSF for 3 d. Dorticos et al (2003) used epidermal growth factor (EGF) in the case of a 15-year-old patient who had HSCT for chronic myeloid leukaemia and developed late-onset HC. EGF was administered at a dose of 2 mg/d into the continuous bladder irrigation. Haematuria settled within 4 d, treatment was continued for a total of 12 d. Both studies had long-term follow up with no recurrence of HC.

Other prevention and treatment approaches

Conditioning and immunosuppressive regimens.  BKV nephropathy (BKVN) causes graft failure in renal transplantation, and type and duration of post-transplant immunosuppression has been recognised as the most important predisposing factor to BKV re-activation (Agha & Brennan, 2006). In this setting, retrospective case series demonstrated that patients who received mycophenolate mofetil (MMF) or Tacrolimus, compared to cyclosporine A, had a higher incidence of BKVN (Mathur et al, 1997; Howell et al, 1999). Pre-emptive reduction or withdrawal of immunosuppression may reduce the incidence of renal allograft failure. A recent randomised controlled trial (Agha et al, 2004) demonstrated that a schedule of early withdrawal of azathioprin and MMF prevented progression of BKV reactivation to BKVN. In the setting of HSCT, however, the situation is more complex as the prevention of GVHD requires a potent immunosuppression. A different approach to this problem is the use of newer immunosuppressive drugs which also have antiviral activity, such as leflunomide (Josephson et al, 2006). In the setting of HSCT, this drug has only anecdotally been used in the management of resistant CMV infection (Avery et al, 2004).

Surgical treatment options

Fibrin glue.  The intravesical application of fibrin glue has been described in a single case report. Purves et al (2005) described the management of late-onset HC following HSCT in an 11-year-old child. The HC progressed to a stage IV with bladder tamponade requiring repeated endoscopic intervention. Active bleeding sites were cauterised and fibrin glue was applied cystoscopically to raw mucosal areas. Haematuria resolved after a single application despite persisting BK viruria. The possible potential complication of this technique is ureteric obstruction secondary to application of fibrin glue to the ureteric orifices.

Selective embolization.  A case report from Gine et al (2003) evaluated the use of selective embolization of the vesical arteries in an adult patient with medical refractory HC secondary to BKV after HSCT. This patient had 45 d of refractory gross haematuria when selective embolisation of the left vesical artery was performed using the ‘Transcatheter embolization technique’ by retrograde catheterization of the femoral artery, super-selective vesical artery catheterisation and embolisation with polyvinyl alcohol microparticles. Haematuria resolved completely within days.

The commonest complication of surgical embolization of the internal iliac artery is gluteal pain. Gangrene of the bladder has also been reported after embolization of the internal iliac artery (Braf & Koontz, 1977; Hietala, 1978). Although a large amount of literature is available on the type of embolization material used for adult intractable hematuria, there is no evidence for the use of embolization in paediatric patients with HC secondary to BKV.

Intravesical hydrostatic pressure.  Intravesical hydrostatic treatment works on the principal of simple tamponade. Bladder blood flow has a linear relationship with bladder pressure (Dunn, 1974). Helmstein (1972) described the application of intravesical hydrostatic pressure (IHP) for the management of bladder cancer.

Intravesical hydrostatic pressure is applied by filling the bladder with normal saline, or introducing a balloon into the bladder which is filled to a pressure approximately 25 cm H2O above the patient’s diastolic blood pressure and left in place for 6 h (Hammonds et al, 1974). Usually control of haematuria can be achieved immediately, but may be short lasting. The temporary tamponade does not cause changes in bladder physiology or serum creatinine (Hansen et al, 1976). A potential complication is rupture of the bladder (partial or total). No literature is available for the paediatric age group.

Cystectomy.  Haemorrhagic cystitis can occasionally be so severe that it not only fails to respond to any conservative measures, but also puts a patient’s life at risk due to uncontrollable haemorrhage or renal failure secondary to complete urinary tract obstruction. Current case literature (Garderet et al, 2000) indicates cystectomy to be the final step in the management of severe medically refractory HC. The authors opted for subtotal cystectomy with urethra and bladder neck preservation, allowing subsequent reconstructive continent urological surgery in the future. The youngest patient to undergo a cystectomy for medical refractory HC was 17 years old.

Whether a radical or subtotal cystectomy should be performed depends largely on the policy of the unit. Andriole et al (Andriole et al, 1990) recommend preservation of the bladder neck in children because severe hemorrhagic cystitis is improved by cystotomy, temprorary urinary diversion and bladder packing. Currently, a subtotal cystectomy with subsequent continent reconstruction is recommended anecdotally, but experience is limited.


Post-engraftment HC can clearly be distinguished from pre-engraftment HC by its clinical and pathobiological features. While pre-engraftment appears to be closely related to the toxic insult of the conditioning regimen, post-engraftment HC is far more complex. A striking feature is the association of post-engraftment HC in timing, incidence and severity with GvHD – suggesting that its pathogenesis may be similarly complex, involving a local inflammatory environment, cellular immune responses, effector mechanisms, and HLA as well as non-HLA genetics. Viewed in a wider context, it seems likely that any measures that reduce the toxic insult of conditioning, prevent GvHD, limit immunosuppression, and promote rapid and functional engraftment will impact on incidence and severity of post-engraftment HC.

The management of the child with HC is difficult. The clinician is confronted with a patient suffering a condition that is potentially life threatening and causing extensive discomfort. The recent toxic insult, profound immunosuppression and co-morbidities, such as renal impairment, severely restrict therapeutic options.

The purpose of this review was to identify effective management options supported by high level scientific evidence as the basis of a clinical guideline. The evidence for interventions most commonly used, and considered ‘conventional’– systemic cidofovir, oestrogen, hyperbaric oxygen therapy, bladder instillation with alum, formalin or prostaglandins – have a weak base of evidence, or are potentially highly toxic, or both.

High quality evidence exists for only two interventions – MESNA and hyperhydration as a preventative measure, and recombinant Factor VII to quell profuse acute haemorrhage. While MESNA/hyperhydration is already considered conventional practice, the use of rFVIIa has little role in the current management of HC other than life threatening acute haemorrhage failing to respond to other measures – its effect is short lasting, it is expensive, and has potential thrombembolic risks.

The limited available evidence highlights the need for more research into the understanding of the pathobiology of HSCT-associated HC. Commonly used interventions with possible clinical benefit (e.g. cidofovir, ciprofloxacin) need to be evaluated in multi-centre, high-quality studies. Potential future preventative and therapeutic options, such as modulation of conditioning, immunsuppression and engraftment, new antiviral and anti-inflammatory and less nephrotoxic agents need to be assessed (Table III).

Table III.   Potential future interventions extrapolated from distinct settings.
InterventionReported contextReference
 Pre-emptive  reduction of immunosuppressionRenal transplantation(Agha et al, 2004)
 LeflunomideRenal transplantation(Josephson et al, 2006)
 OrgoteinRadiation cystitis(Kadrnka, 1981; Marberger et al, 1981; Sanchiz et al, 1996)
 Sodium pentosan polysulphateInterstitial cystitis, radiation cystitis(Parsons, 1986; Hampson & Woodhouse, 1994; Hwang et al, 1997; Sandhu et al, 2004)
 BerberineAnimal study(Xu & Malave, 2001)
 N-acetylcysteineAnimal study(Palma et al, 1986)
 Εpsilon-aminocaproic acidHaemorrhage after prostrate resection,
cystitis after radiation/chemotherapy
(Sharifi et al, 1986; Singh & Laungani, 1992)
 Sairei-toRadiation/non-radiation cystitis(Shida et al, 1994)
 OrgoteinRadiation cystitis(Kadrnka, 1981; Marberger et al, 1981; Sanchiz et al, 1996)
 Sodium pentosan polysulphateInterstitial cystitis, radiation cystitis(Parsons, 1986; Hampson & Woodhouse, 1994; Hwang et al, 1997; Sandhu et al, 2004)
 NystatinIn vitro(Roskopf et al, 2006)
 Retinoic acidsIn vitro(Roskopf et al, 2006)


But where does this leave the clinician having to decide how to manage the patient with post-engraftment HC? In discussion we identified four guiding principles, which we found reflected in our clinical practice (Table IV):

Table IV.   Implications for clinical practice.
Implications for clinical practice
Reduction of risk factors for post-engraftment HC
Optimal supportive treatment
Close cooperation between physicians and surgeons needed in optimal management of HC
‘Conservative’ approach avoiding interventions with long-term consequences, as post-engraftment is a transient condition
Cautious application of any interventions other than supportive measures: evidence for interventions regarded as ‘conventional’ is largely anecdotal, and most interventions are potentially harmful
  • • Prevention by addressing known risk factors early: best possible donor-recipient matching, using the least toxic conditioning regimen possible with MESNA/hyperhydration, tight monitoring of viral titres and prompt treatment of re-activation in the peri-transplant period, GvHD prevention and tightly monitored immunosuppression.
  • • Optimal supportive treatment of manifest HC, as a primum non-nocere approach and accompanying further management: ensuring appropriate hydration and renal function, haematological homoeostasis (maintaining high platelet counts, appropriate red cell counts and levels of clotting factors), pain relief, catheterisation with cystoscopic clot extraction and continuous bladder irrigation with normal saline for prevention of clots and bladder tamponade, if necessary (Table V).
  • • Early and close collaboration between medical and surgical teams in the management of these patients to coordinate and optimise timing of interventions.
  • • As post-engraftment HC is by nature a transient condition that resolves with immune reconstitution – we would aim for a ‘conservative’ approach avoiding measures that may inflict long-term consequences on the patient. Given the low grades of recommendation, any further interventions would have to be considered on an individual basis for a given clinical scenario, carefully balancing benefits and risks.
Table V.   Supportive treatment of HC.
MorphinePain controliv/im/intravesical(Duckett et al, 1997, McCoubrie & Jeffrey, 2003; BNF 2006)
OxybutininPain, spasms, urgencypo/intravesical(Bemelmans et al, 2000; Lehtoranta et al, 2002, BNF 2006; Kilic et al, 2006)
TolterodinePain, spasms, urgencypo(Goessl et al, 2000, BNF 2006; Kilic et al, 2006)
Potassium citrateUrine alkalinisationpo(BNF 2006)
Catheterisation, bladder irrigation with 0·9% saline or 3% sodium citrate≥Grade II HCintravesical(BNF 2006)
Lidocaine/bupivacainePain controlintravesical(Clapp et al, 1999; Chiang et al, 2005)
Substitution with blood productsAnaemia, thrombocytopenia, clotting factor deficiencyiv 


While the advice and information in these guidelines is believed to be true and accurate at the time of going to press, neither the authors, the British Society for Haematology nor the publishers accept any legal responsibility for the content of these guidelines.


The authors thank Dr Peter G Middleton, Senior Lecturer in Experimental Haematology at Newcastle University, for constructive feedback on several draft versions of the manuscript.


Appendix I: Evidence levels and grading of recommendations (Harbour & Miller, 2001)

  1. RCT, randomized controlled trial.

Levels of evidence
 1++High quality meta-analyses, systematic reviews of RCTs, or RCTs with a very low risk of bias
 1+Well conducted meta-analyses, systematic reviews of RCTs, or RCTs with a low risk of bias
 1−Meta-analyses, systematic reviews or RCTs, or RCTs with a high risk of bias
 2++High quality systematic reviews of case-control or cohort studies or
High quality case-control or cohort studies with a very low risk of confounding, bias, or chance and a high probability that the relationship is causal
 2+Well conducted case-control or cohort studies with a low risk of confounding, bias, or chance and a moderate probability that the relationship is causal
 2−Case-control or cohort studies with a high risk of confounding, bias, or chance and a significant risk that the relationship is not causal
 3Non-analytic studies, eg case reports, case series
 4Expert opinion
Grades of Recommendation
 AAt least one meta-analysis, systematic review, or RCT rated as 1++ and directly applicable to the target population or A systematic review of RCTs or a body of evidence consisting principally of studies rated as 1+ directly applicable to the target population and demonstrating overall consistency of results (can be trusted to guide practice)
 BA body of evidence including studies rated as 2++ directly applicable to the target population and demonstrating overall consistency of results or Extrapolated evidence from studies rated as 1++ or 1+ (can be trusted to guide practice in most situations)
 CA body of evidence including studies rated as 2+ directly applicable to the target population and demonstrating overall consistency of results or Extrapolated evidence from studies rated as 2++ (provides some support for recommendations but care should be taken in its application)
 DEvidence level 3 or 4 or Extrapolated evidence from studies rated as 2+ (evidence is weak and recommendation must be applied with caution)