Pyoderma gangrenosum following an unplanned caesarean section: a patient revisited
- Waseem Ghumra ,
- Adam Gold and
- Richard Michael Azurdia
- Dermatology, Broadgreen Hospital, Liverpool, Liverpool, UK
- Correspondence to Dr Waseem Ghumra; w.ghumra@nhs.net
Abstract
A 35-year-old woman was referred urgently to the dermatology department because of significant wound breakdown 3 weeks following an emergency caesarean section. Examination revealed a full thickness, undermined ulcer spanning the width of the patient’s caesarean scar, exposing the patient’s uterus. Clinical appearances were consistent with pyoderma gangrenosum. The patient has a history of complicated pyoderma gangrenosum, having undergone skin grafts 14 years prior, for ulcers on her lower legs. That episode was the subject of a case report, published in the BMJ, in 2007. On this occasion, the patient was treated with oral corticosteroids, and ciclosporin based on its efficacy during her previous episode, which in conjunction with negative pressure wound therapy, resulted in complete re-epithelialisation of her ulcer within 6 months.
Background
Pyoderma gangrenosum (PG) is an uncommon neutrophilic dermatosis. It classically presents as a rapidly progressive and painful ulcer with a violaceous, undermined border.1 It can affect patients of any sex or age, but has a predilection for females and has a median age at presentation of 59.2 In approximately half of all cases, it is associated with another comorbidity, most commonly inflammatory bowel disease, rheumatoid arthritis and myeloid dyscrasias.3 It can affect any site, and classically presents on the leg, although diagnosis is often delayed, as PG can mimic other ulcerative conditions, and is a diagnosis of exclusion.1 Treatment of PG can be challenging. There are no national guidelines to direct management and only a single randomised controlled study comparing treatments.4 As with most inflammatory dermatoses, management is largely immunosuppressive, with high dose oral steroids being the mainstay of treatment.5 In the longer term, ciclosporin, mycophenolate mofetil or dapsone can be used. Successful outcomes have also been reported with biological medications such as infliximab, adalimumab, etanercept and ustekinumab.6 Surgical debridement is under most circumstances, contraindicated because of the poorly understood phenomenon of pathergy; this is an exaggerated inflammatory response to injury and is responsible for the de novo appearance of PG postoperatively in susceptible patients.7
Pregnancy is thought to be protective in some inflammatory skin conditions such as plaque psoriasis although generally, the pregnant state may worsen existing skin pathologies or precipitate new ones.8 9 The possible reasons for this are discussed below. The specific relationship between PG and pregnancy is not clear. The vast majority of cases of PG in pregnancy, however, tends to develop in latter half of pregnancy, or following a caesarean section.10 Treatment of PG in this group is challenging, not least because of the effects of steroids and more importantly, steroid-sparing agents on the unborn child; and when postpartum, the breastfeeding mother.
Case presentation
A 35-year-old female patient of gravida 2 para 1 underwent an emergency caesarean section at 31 weeks following a placental abruption. She presented 6 days following the procedure, with significant abdominal pain and was noted to have a violaceous border around her caesarean site. She was commenced on topical steroids by the gynaecology team, to treat presumed early PG and intravenous Augmentin and doxycycline, to cover for a potential surgical site infection. Oral prednisolone was added when there was no demonstrable improvement.
Despite these measures, the patient’s caesarean site continued to deteriorate. On presentation to the dermatology department approximately 3 weeks postcaesarean section, examination of the patient’s lower abdominal wall revealed a full-thickness defect, exposing her uterus and measuring 24 cm across by 5 cm craniocaudally (figure 1). The edge of the ulcer was undermined. Although significant slough was present, there was no obvious secondary infection and a bacterial swab showed contaminant growth only.
The patient’s abdominal wound, on presentation to the dermatology department.
The patient had given birth by normal vaginal delivery 2 years prior. She had sustained a minor vaginal tear which did not provoke any pathergic response. Of note, the patient also had a laparoscopic appendicectomy during the first trimester of this latest pregnancy, which again, was complication-free.
Approximately 14 years ago, the patient presented to the emergency department with a painful blister on her left lower leg.11 This was thought to be an abscess and she underwent a series of surgical debridements and subsequent skin grafting, culminating in an admission to the intensive care unit. Dermatology opinion was then sought, and a diagnosis of PG was made. The patient was initiated on intravenous methyl prednisolone, and subsequently, ciclosporin 300 mg once daily, which was switched to mycophenolate mofetil 3 g once daily in combination with prednisolone 40 mg once daily.
Investigations
The histopathology of PG is non-specific and biopsies are normally carried out to exclude other causes of ulceration. A biopsy may show a neutrophilic infiltrate with secondary leucocytoclastic vasculitis and in later lesions, necrosis of the skin.12
In this case, because of the patient’s preceding history of PG, a biopsy was not carried out. A deep incisional biopsy would also have potentially resulted in further break down of the abdominal wound.
The patient’s C reactive protein (CRP) and white cell count peaked at 416 (normal value <5) and 30.3 (range 3.5–11), respectively, 7 days following her caesarean section, with the CRP normalising approximately 2 weeks after.
Serial swabs of the patient’s wound bed were negative; wound swabs in PG will typically not show organism growth or show contaminant growth only.7
Differential diagnosis
The main differential diagnosis in this case would be a surgical site infection±pelvic suppurative collection. A CT abdomen with contrast was requested, demonstrating a small collection/haematoma, which was not felt to be clinically significant.
Treatment
The size of the patient’s defect, as well as her postpuerperic state made this a complex medical presentation. Negative pressure wound therapy (NPWT) was initiated after consulting the dermatology literature on the management of similarly significant defects in patients with PG. With the support of the tissue viability team, the patient was exposed to an NPWT system (figure 2), with the pressure set at ~125 mm Hg, on continuous therapy, for a total of 3 months.
Negative pressure wound therapy setup.
The patient’s prednisolone was increased to 60 mg once daily, which was weaned completely over the next 2 months. As the patient was not breast feeding, the decision was made to commence ciclosporin, which was introduced at a dose of 150 mg two times per day (3 mg/kg), reduced to 50 mg two times per day after 5 months, and stopped completely after 6 months, after complete re-epithelialisation was achieved.
Outcome and follow-up
The patient attended the dermatology department twice weekly for reapplication of her negative pressure dressing, for a total of 3 months, by which time her abdominal wall defect had halved in width, measuring 11.5 × 1.2 cm with healthy granulation tissue evident (figure 3). Further dressing of the patient’s wound was carried out in the community under the care of the district nurses. The patient continued to be seen in the outpatient department for a total of 6 months following her initial presentation. At the time of her discharge, her wound had completely healed (figure 4).
The patient’s abdominal wound after 3 months of negative pressure wound therapy
The patient’s abdominal wound, at approximately 6 months
Discussion
A favourable course in pregnancy is dependent on the maintenance of immunological tolerance between mother and fetus. This involves a complex interplay of various cell signalling pathways, with particular upregulation of anti-inflammatory cytokines such at Transforming Growth Factor-Beta (TGF-B) and interleukin-10 (IL-10), before proinflammatory pathways predominate, prior to parturition.13 It is not known how the cytokine profile changes in pregnancy in patients with PG or with a propensity to develop PG, relative to unaffected controls. In all PG patients, the cytokine profile is shifted towards a T-Helper 1 and T-Helper 17 (TH1/TH17) bias.14 It is not unreasonable to assume then that in patients who develop PG in pregnancy, there is an aberrant proinflammatory shift that gives rise to a protracted inflammatory response, a cardinal feature of PG. This may partially explain why emergency presentations such as placental abruption and chorioamnioitis are disproportionately seen in this cohort of patients, and would fit with previously published reports demonstrating higher levels of proinflammatory cytokines in these pathologies.10 15 16
PG is a recognised but rare complication of a caesarean section, with 20 cases described in the English language literature. In five of these cases, the disease occurred in patients with a previous history of PG, and so was not unanticipated.17–21 Where the condition manifests de novo in surgical incision sites, however, wound infections are almost always suspected.7 While reasonable as an initial differential, the delay in diagnosis inevitably results in prolonged courses of unnecessary antibiotics or painful surgical debridement. Our review of these 20 cases of postcaesarean PG demonstrated that a surgical site infection or necrotising fasciitis was the initial diagnosis in 18 patients (table 1). The majority of these cases exhibited negative wound cultures. Additionally, in 9 of the 20 cases, the patients underwent at least one surgical debridement prior to a diagnosis of PG being made.
A summary of previously published cases of post-caesarean pyoderma gangrenosum
| Patient age | Gestational age | Factors complicating pregnancy | Initial diagnosis and wound culture results | Treatment |
| 34 (27) | 36 | Premature rupture of membranes | Surgical site infection. Negative wound culture | Corticosteroids |
| 32 (28) | Not stated | Not stated, presumed elective | Necrotising fasciitis, patient underwent surgical debridement | Corticosteroids, intravenous immunoglobulins |
| 21 (29) | 32 | Preterm labour, breech | Surgical site infection and sepsis. Negative wound culture. Patient underwent surgical debridement. | Corticosteroids |
| 29 (17) | 37 | Elective | Surgical site infection. Negative wound culture | Corticosteroids |
| 27 (17) | 32 | Chorioamnionitis | Surgical site infection. Negative wound culture | Corticosteroids |
| 34 (30) | 34 | Premature rupture of membranes | Surgical site infection. Negative wound culture. Patient underwent surgical debridement | Corticosteroids |
| 32 (24) | >7 months | Chorioamnionitis | Surgical site infection. Negative wound culture. Patient underwent surgical debridement | Corticosteroids. Immunoglobulins, skin graft, VAC dressings and Cyclosporine |
| 27 (18) | 27 | Chorioamnionitis | Surgical site infection. Swab results not documented | Corticosteroids |
| 28 (31) | Not stated | Not stated | Necrotising fasciitis. Wound culture initially positive but further cultures negative. Patient underwent surgical debridement | Corticosteroids and Dapsone |
| 23 (32) | Not stated | Placental abruption | Surgical site infection. Staph epidermidis isolated from swab. Patient underwent surgical debridement | Corticosteroids |
| 39 (33) | 36 | Placenta previa. Placenta accrete, vaginal bleeding | PG as initial diagnosis. Wound cultures negative | Corticosteroids |
| 33 (34) | Term | Premature rupture of membranes, placenta previa | Surgical site infection. Swab results not documented | Corticosteroids |
| 22 (35) | Not stated | Elective, breech | Surgical site infection. Negative wound culture. Patient underwent surgical debridement | Corticosteroids, Dapsone, and Cyclosporine |
| 32 (19) | 37 | Not stated | Surgical site infection. Swab positive for Pseudomonas aeruginosa and methicillin-resistant coagulase-negative Staphylococcus | Corticosteroids, skin graft, VAC therapy |
| 32 (20) | 32 | Placental abruption | Surgical site infection. Swab results not documented | Corticosteroids and Dapsone |
| 32 (36) | 31 | Foetal distress | PG as initial diagnosis | Corticosteroids |
| 29 (37) | 37 | Elective | Surgical site infection. Negative wound culture | Corticosteroids and Ciclosporin |
| 36 (38) | 39 | Elective | Necrotising fasciitis, patient underwent surgical debridement. Tissue swab not ducmented, tissue culture negative | Corticosteroids and Ciclosporin |
| 35 (21) | Not stated | Cephalopelvic disproportion | Surgical site infection. Negative wound culture. Patient underwent surgical debridement. Negative wound culture | Corticosteroids and Ciclosporin |
| 36 (39) | 29 | Foetal distress | Surgical site infection. No documentation on swab results. | Corticosteroids followed by debridement and skin grafting |
-
PG, pyoderma gangrenosum.
NPWT is used routinely following large surgical procedures to aid healing and to reduce the frequency of secondary surgical site infections.22 A negative pressure system was first used in the management of PG in 2005 but is not commonly seen in the management of this condition in the UK.23 In the context of postcaesarean PG, NPWT has been used twice previously, as far as we are aware, although in both cases, the patients also underwent skin grafting.19 24
The role of prophylactic systemic corticosteroids in patients with a history of PG undergoing surgical procedures is unclear. Ohmaru-Nakanishi et al report a case of a patient who developed perineal PG during her third trimester.25 She was commenced on oral prednisolone and underwent a caesarean section. She did not go on to develop PG affecting her caesarean incision site. Several authors report positive outcomes with perioperative steroids in patients undergoing general surgical procedures. Others, however, report that their patients developed PG despite perioperative steroids.20 Canzoneri et al describe the case of a young woman with a history of PG, who underwent multiple complication-free surgeries over a 10-year period without immunosuppression.26 Clearly, there is no consensus regarding the benefits of prophylactic systemic steroids in patients in this cohort. At present, we feel that there is insufficient evidence to justify their use, especially in women undergoing a caesarean section.
Postcaesarean PG is associated with considerable morbidity. This is especially the case if the patient undergoes surgical debridement, as further wound breakdown is likely, via a pathergic response. Features that favour PG over a surgical site infection include a previous history of PG, negative wound/blood cultures and a comorbidity classically linked with PG. Strikingly, we also note that postcaesarean PG is disproportionately associated with emergency caesarean section presentations, and relate this to abnormal pro-inflammatory signalling although the mechanism is unclear. We strongly suggest a multidisciplinary approach is adopted when patients with presumed postcaesarean surgical site infections fail to respond appropriately to antimicrobial therapy.
Patient’s perspective
Prior to having children, I was aware that there was a risk to the PG flaring up following any c-section as I had previously experienced PG approximately 14 years ago. My first episode of PG led to 3 months as an in-patient at hospital, narrowly avoiding having my right leg amputated and losing most of the skin on my right leg and some of the skin on my left leg.
Having delivered one baby vaginally, I did not anticipate that my second son would be delivered by c-section. When he was, I was worried as I knew that there was a risk of the PG returning and spent the initial period after his deliver anxiously waiting for any symptoms to appear. It took approximately 3 days after the c-section for PG symptoms to start. Due to the pain associated with PG it was apparent to me that the wound was breaking down due to PG, but it took some days for the midwives and Drs who were looking after the delivery of my baby to administer steroids to dampen down the immune system response. The period between PG symptoms starting and the first dose of steroids being administered (approximately 6 days after delivery) was painful and difficult as I was also trying to care for a new baby, born 5 weeks early.
Once the immunosuppressant drugs were started, the wound stopped deteriorating and I felt an improvement in my symptoms. Once my son was discharged from the neo-natal unit and I could be treated by dermatology (based at a different hospital), I felt an improvement every week. The most significant improvement in my care, compared to when I first had PG, was that this time I managed to avoid a significant period as an in-patient at hospital (and all of the risks that brings with regard to infection etc), due to the vacuum wound care system. The vac system also had a significant, positive impact on my pain which was brilliant. Most importantly this meant that I could be at home with my family and with our new baby.
Learning points
-
Pyoderma gangrenosum can mimic surgical site infections and should be considered in patients with postcaesarean incisions that exhibit delayed healing, or that do not respond to antibiotics.
-
The benefit of prophylactic corticosteroid use is unclear in this cohort of patients and the clinician should weigh the risk of the development of pyoderma gangrenosum against the possibility of peurpural sepsis. Dermatology advice should be sought before a planned caesarean section in patients with a history of pyoderma gangrenosum.
-
Negative pressure wound therapy is a useful adjunct in the management of postcaesarean pyoderma gangrenosum.
Acknowledgments
We would like to thank the Tissue Viability Nurses, particularly Natalie Murphy and Lisa Robson, based at the Royal Liverpool University Hospital, for their invaluable expertise and input with the negative pressure wound therapy system.We would also like to thank Angela Hall, Manager at the Liverpool University Hospital Foundation Trust Library, for obtaining access to many of the articles cited in the submissionFinally, we would like to thank the Photography Department at the Royal Liverpool Hospital.
Footnotes
-
Contributors The patient was seen by myself on-call. RMA was the patient’s named consultant, and the consultant under whose care the patient was managed for the duration of her treatment. RMA approached the patient to enquire whether she would consent to being the subject of a published case report. The planning of the article was undertaken by myself (WG), AG and RMA; the content of the article was largely decided beforehand. While the majority of the written piece for submission was drafted by myself, all authors contributed significantly to the body of the text, and proof-read the final version for submission.
-
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
-
Competing interests None declared.
-
Patient consent for publication Obtained.
-
Provenance and peer review Not commissioned; externally peer reviewed.
- © BMJ Publishing Group Limited 2021. No commercial re-use. See rights and permissions. Published by BMJ.
References
Use of this content is subject to our disclaimer