Purpura fulminans is an infrequent but potentially catastrophic illness. It is defined as a severe and often fatal form of idiopathic thrombocytopenia purpura that occurs mostly in children following a meningococcal infection, although it may occur with other bacterial infections. The mortality rate associated with purpura fulminans is high, with most deaths occurring within forty-eight hours after onset. Patients who survive the initial acute phase of fulminant meningococcemia are at increased risk for serious complications as a result of poor tissue perfusion1. Extensive soft-tissue loss, auto-amputation, and surgical amputation due to extremity infarction and gangrene may occur. In the initial stage, it may be difficult to determine the degree of tissue involvement. However, performance of skin and soft-tissue-releasing incisions of the necrotic tissue to reduce the prevalence of extremity necrosis is controversial and probably not recommended, so it is rare that early surgical intervention is required2,3. Once the patient is stable and the demarcation between necrotic and viable tissue becomes more clearly established, debridement of all necrotic tissue is essential and may necessitate extensive removal of skin, subcutaneous tissue, muscle, and bone4.
Following extremity amputations and soft-tissue healing, major issues with the residual extremities often persist. Stump overgrowth, growth disturbances, and scar contractures are common results, but they have been infrequently addressed in the orthopaedic literature4-11.
The aim of this study was to describe our experience with purpura fulminans related to meningococcemia, with an emphasis on long-term sequelae and their treatment.
This study, which was approved by our institutional review board, included children with a primary diagnosis of meningococcal sepsis who were referred to our institution and eventually required orthopaedic management for musculoskeletal problems.
A retrospective chart and radiograph review was performed for all patients with a history of meningococcal septicemia treated at our institution between 1987 and 2008. Cases were identified by the examination of charts coded as meningococcal septicemia by the health records department and were verified by chart review. Medical charts were reviewed to determine patient and meningococcal lesion characteristics, associated symptoms, surgical management, and evidence of postinfectious complications. The inclusion criterion for this study was confirmation of meningococcal infection by a positive culture of blood, cerebrospinal fluid, or a skin lesion in patients who were referred to our institution for orthopaedic evaluation and management related to the disease. We followed patients who had survived the acute phase of the disease and developed musculoskeletal sequelae. Patients under twenty-one years of age were regularly followed at our outpatient children's amputee clinic, and both a clinical examination and a retrospective chart and image review were performed for the purposes of this study. Patients older than twenty-one years of age were no longer followed at our institution, and only a retrospective chart and image review was performed for the purpose of the present study.
Medical records were specifically reviewed to determine the age at the onset of the sepsis, age at surgery, number of surgical procedures, and musculoskeletal sequelae. Radiographs, magnetic resonance images, and computed tomography scans, when available, were also reviewed to determine the extent and location of growth abnormalities.
Musculoskeletal complications were classified according to the upper and lower-extremity segments that were involved and according to the type of sequelae. Sequelae were characterized into two main groups: early and late. Early musculoskeletal sequelae required pediatric orthopaedic treatment within six months after the onset of sepsis and were primarily amputations. Late sequelae required surgical management six months or more following the initial infection. Late complications included growth disturbances, stump overgrowth, scar contractures, and soft-tissue and bone infections.
Source of Funding
There was no external funding source for this study.
Forty-eight patients met the inclusion criteria for the study. There were twenty-two boys and twenty-six girls. The mean age (and standard deviation) at the time of the onset of the meningococcal sepsis was 2.6 ± 3.0 years (range, 0.2 to 13.4 years). At the time of chart and radiograph review, the average age of this patient group was 18.6 ± 6.9 years (range, 1.6 to 33.2 years), and the mean duration of follow-up was 11.7 ± 5.1 years (range, one to twenty years).
The total number of surgical procedures was 212, and the average number was 4.4 ± 2.9 (range, one to twelve) per patient. Twenty-two children (46%) required three surgical procedures or fewer; eighteen (38%) had four, five, or six; and eight (17%) had seven or more.
Early Orthopaedic Sequelae
Early sequelae included amputations at a variety of levels. Twenty-one (44%) of the forty-eight patients had an amputation of at least one upper-extremity segment. Nine of these twenty-one patients had at least one major upper-extremity amputation, defined as through or proximal to the wrist. Thirty-four (71%) of the forty-eight patients had an amputation of at least one lower-extremity segment. Twenty-four of these thirty-four had at least one major lower-extremity amputation, defined as either a below-the-knee amputation or a more proximal amputation. Twelve (25%) of the forty-eight patients sustained four-extremity amputations. Of these twelve children, six had amputations proximal to both wrists and both feet. Details of the extremity amputations are shown in Table I.
Late Orthopaedic Sequelae
Late sequelae included growth disturbances, stump overgrowth, scar contractures, and infections. It is important to note that thirteen (27%) of the forty-eight patients had developmental delay at the time of follow-up, and it was severe in seven of them. Four of these seven patients had central involvement due to brain infarction; they became nonambulatory and noncommunicating and had spastic quadriplegia (Gross Motor Function Classification System [GMFCS] level IV or V12).
Growth Disturbances
Growth disturbances occurred in both the upper and the lower extremities. Five (10%) of the forty-eight patients developed growth disturbances in the physes of the upper extremity. All angular deformities of the upper extremities occurred at the distal part of the forearm in limbs without an amputation and resulted in radial deviation due to distal radial growth disturbance and relative ulnar overgrowth (four patients). One patient with a below-the-elbow amputation developed a complete growth arrest of the distal part of the humerus and the proximal, residual part of the forearm (radius and ulna), leading to a very short stump without angular deformity. Growth disturbances leading to angular deformities were less common in the upper extremity than they were in the lower extremity.
Twenty-six (54%) of the forty-eight patients developed lower-extremity growth disturbances, which were often bilateral, with sixty lower-extremity physes involved. The most commonly involved physes were in the proximal parts of the tibia and fibula (n = 32; 53%), the distal part of the femur (n = 13; 22%), the distal parts of the tibia and fibula (n = 12; 20%), and the proximal part of the femur (n = 3; 5%).
A total of thirty-six angular deformities were identified. Although the deformities can occur along any direction of the axial plane, many can be classified as a predominantly coronal or sagittal plane deformity. The most common coronal plane deformities were knee varus (n = 18; 50%), knee valgus (n = 5; 14%), ankle varus (n = 2; 6%), ankle valgus (n = 1; 3%), and proximal femoral varus (n = 1; 3%). The most common sagittal plane deformities were recurvatum of the knee (n = 5; 14%) and procurvatum of the knee (n = 4; 11%), with the majority being proximal tibial deformities. Multiplanar deformities were very common. Angular deformities ranged in magnitude from moderate to severe and increased with age. As a rule, the shorter the residual limb distal to the involved physis, the longer the delay in treatment as the angular deformity can be accommodated in a prosthesis. In our series, the greatest deformity before surgical intervention was 65° (Figs. 1-A, 1-B, and 1-C). The deformities in the plane of the range of motion are better tolerated and better compensated for by prosthetics than are the ones perpendicular to it. Table II lists the angular deformities by extremity.
Stump Overgrowth
There was no stump overgrowth in any of the twenty-one patients with upper-extremity amputations. Stump overgrowth was present in eleven (32%) of the thirty-four patients with lower-extremity amputations. Table III summarizes late orthopaedic sequelae of meningococcal septicemia, including stump overgrowth.
Scar Contracture
Scar contractures occurred in both the upper and the lower extremities. Fourteen (29%) of the forty-eight patients had upper-extremity scar contracture and nineteen (40%) of the forty-eight had lower-extremity scar contracture. The prevalence of scar contractures was similar in the upper and lower extremities. Fifty scar contractures required surgical release or revision; the locations of these releases or revisions included the knee (n = 18; 36%), the foot (n = 8; 16%), the hand and wrist (n = 7; 14%), the forearm (n = 3; 6%), the axilla (n = 3; 6%), the elbow (n = 2; 4%), the thigh (n = 2; 4%), the calf (n = 2; 4%), the buttock (n = 2; 4%), the arm (n = 2; 4%), and the groin (n = 1; 2%).
Knee surgical procedures included scar revision alone (six cases), scar revision and skin grafting (three cases), distal femoral extension osteotomy (three cases), z-plasty (two cases), scar revision with hamstring release and skin grafting (one case), capsulotomy with hamstring lengthening and skin grafting (one case), scar revision and capsulotomy (one case), and scar revision after implantation of a soft-tissue expander (one case). Repeated release was necessary for seven knee scars (in six patients) and the rate of repeat surgery was 40%.
Foot contractures required scar revision alone (one); scar revision with metatarsal head resection (one); soft-tissue release with skin-grafting (one); plantar release, Achilles tendon lengthening and tibialis posterior transfer, and scar release (one); z-plasty with extensor hallucis longus lengthening (one); scar revision with Achilles tendon and tibialis posterior tendon lengthening (one); or scar revision with Achilles tendon lengthening (two).
Axilla contractures required z-plasty (one) or surgical scar revision (two). Repeat surgery was necessary in one case.
Hand and wrist surgical procedures included scar revision alone (two), z-plasty with palmaris longus release (one), soft-tissue release with flap closure (two), and deepening of web spaces (two). Repeat surgery was necessary for one hand scar.
Elbow contractures required z-plasty with skin-grafting (two). Forearm (three) as well as arm (two) scars required surgery for scar revision alone. Calf surgery (two) consisted of skin grafting after scar release and revision. Buttock (two), thigh (two), and groin (one) contractures required scar revision alone. Table III summarizes late orthopaedic sequelae of meningococcal septicemia, including scar contractures.
Soft-Tissue and Bone Infections
Ten (21%) of the forty-eight patients acquired a soft-tissue and bone infection at a mean of 7.7 ± 5.7 years (range, one to sixteen years) following treatment for meningococcal sepsis. Eight patients developed the infection in the distal portion of the amputation stump (after a below-the-knee amputation in six, after an above-the-knee amputation in one, and after a Syme amputation in one). Of the remaining two patients, one had the infection in the proximal part of the tibia of an intact limb and the other had it in the thigh of an intact limb. Table IV shows the microorganisms that were isolated and the sites of infection.
Overall, there are few reports describing the surgical musculoskeletal management and outcomes of children with a history of meningococcal sepsis. In a recent study on the management and outcomes of 143 patients diagnosed with meningococcal sepsis, Bache and Torode5 reported a 14.7% mortality rate. Of the patients surviving the acute phase of the disease, one-third developed late musculoskeletal sequelae.
In recent years, probably because of prompt diagnosis and effective aggressive resuscitation, the majority of children with fulminant meningococcemia have survived. Most of the patients included in the present study had a purpuric skin rash that progressed to multiple areas of cutaneous gangrene associated with extensive necrosis of underlying subcutaneous fat, fascia, muscle, and even bone. Tissue necrosis was most extensive in regions of reduced blood flow, such as the extremities, resulting in multiple amputations that involved all four limbs of some patients.
Late-onset musculoskeletal sequelae have been reported primarily in case studies, but there is a lack of comprehensive reviews of the subject8,10,11,13,14. In our study, late musculoskeletal sequelae were more common than expected and involved both limbs with and those without an amputation. Several studies suggest that musculoskeletal problems are associated with more extensive tissue damage5,6,10,11. In our series, in which the mean duration of follow-up was over eleven years, growth disturbances tended to occur more frequently in limbs without an amputation than they did in those with an amputation. Although the pathology of this association is not clearly understood, our data suggest that growth plates both beneath areas of skin necrosis5 and in limbs free of amputations are at risk for growth arrest. The physis may not appear to be involved early on, but with time the insult may become evident. It is likely that angular deformities are more obvious in longer-segment limbs. Minor angular deformities in the short segment of an extremity post-amputation are easily accommodated in a prosthetic socket and less likely to be diagnosed or require surgical treatment. It is also possible that some of the extremities with an amputation had infarction of the whole growth plate or at least a major portion of it, resulting in complete growth arrest rather than angular deformity. If this was the case, these growth arrests were not diagnosed, particularly in patients with involvement of multiple extremities, as radiographs were made only for patients with clinical issues.
All osseous bridges and physeal abnormalities were seen on conventional radiographs in this study. Conventional tomography was used for several patients with the longest follow-up, but this technique was abandoned in the mid-1990s and was replaced by magnetic resonance imaging and computed tomography. Those modalities can more accurately depict and detail the location, size, shape, and contour of the growth plate and the osseous lesion, which is crucial for planning surgical treatment. Magnetic resonance imaging is more accurate than conventional techniques as it has multiplanar capabilities to allow cross-referencing of images in two planes and facilitate physeal mapping. Magnetic resonance imaging has excellent contrast and spatial resolution and has the advantage of detecting fibrous and cartilaginous bars, as demonstrated experimentally and clinically15-18. Our experience has revealed that any part of the physis, including the central part, is vulnerable to the meningococcal infection. The sections of the physis that are more likely to be affected depend on which physis is involved. Our findings confirm the previously reported observation that, at the distal part of the tibia, the anteromedial section of the physis is most susceptible to growth arrest, whereas the proximal part of the femur and the proximal part of the tibia are more susceptible to physeal bars in their periphery7,15-18.
Multiple high-level amputations were common in our series. Twelve (25%) of our patients had amputations affecting all four limbs, with six of them sustaining amputations proximal to both wrists and both feet. Proximal amputations in the upper extremities occurred less frequently than did proximal amputations in the lower extremities in our study population. This finding corroborates previous reports in which extremity involvement was much more often peripheral than it was central and in which there was a predilection for the lower extremities6,14-17. Moreover, the prevalence of patients with zero or one amputation was much lower than the prevalence of patients with two or more (Table I). This finding is in contrast to that in the series of Bache and Torode, who reported low frequencies of proximal amputations or multiple amputations5. Bache and Torode reviewed the records of all patients admitted to their institution with a diagnosis of meningococcal septicemia. They subsequently identified patients with acute and chronic vascular and skeletal complications and assessed the results of surgical interventions. Our study, on the other hand, focused on patients who had survived the acute phase of the disease and developed musculoskeletal sequelae. There may be other possible explanations for the difference in the amputation levels between the two studies, with one being that the patients in the two studies differed with regard to their chances of surviving the acute episode. It is possible that some of the patients with the highest-level amputations in our study were those with more severe involvement by the meningococcal sepsis and that such patients did not survive in the previously reported study. As there were no data on mortality in our study and the severity of the original acute episode was not classified in either study, this explanation is only theoretical. In the experience of the senior author (J.I.K.), amputation proximal to the midpart of the tibia has been rarely required if the course of the disease was followed expectantly and amputation was performed when both superficial and deep-tissue necrosis levels were clearly demarcated19.
Scar contracture leading to joint contracture and loss of joint movement is an important and relatively frequent finding18,20-23. The upper and lower extremities tend to have similar prevalences of scar contractures. Joint contractures appear to be the factor that most limits the progression to full activity and weight-bearing by patients requiring a prosthesis. In our series, the most common location of scar contractures requiring surgical release to restore adequate joint motion was at the level of the knee. Surgical release and adequate physical therapy are often necessary to restore motion for an acceptable long-term outcome. However, recurrence of the contracture is possible and repeated surgery is often needed. Although our numbers were not large enough for us to make definitive conclusions, it does not appear that younger patients had fewer recurrences and a better gain of joint motion.
Patients with long-term sequelae of meningococcal sepsis are at increased risk of developing soft-tissue and bone infections. On the average, infections developed several years after the onset of the meningococcemia, and they were always located in the lower extremity, mostly in patients with a below-the-knee amputation. We believe that this is the first report to document local soft-tissue and bone infections in patients who survived meningococcal sepsis. We hypothesize that the relatively high prevalence of soft-tissue and bone infections is probably related to the fact that most patients required skin grafts and developed retractile and keloid scars resulting in poor, fragile, and uneven skin coverage. This results in skin folds of different depth and shape, making hygiene and skin care difficult. Moreover, the uneven surface makes prosthetic fitting a challenge. The poor skin coverage in association with an imperfect prosthetic fit can predispose to skin damage, which becomes a potential entry point for bacteria.
Custom prosthetic fitting is required for this patient population. A detailed analysis of prosthetic fitting was not part of this study but, anecdotally, the silicone sleeve socket was best for our patients.
Decreased vascularity of the scar tissue and bone of the stump may compromise further biologic defenses against pathogens. In our patients, the secondary infection usually resolved following surgical debridement and antibiotic therapy. However, four of the ten patients with secondary infection had a recurrence of that infection at the same anatomical site.
The late complication of stump overgrowth and infection is primarily associated with transosseous amputations in children, which can be avoided in the majority of cases. Technetium-99 bone scintigraphy is a useful adjunct to clinical assessment in differentiating viable from nonviable tissues in patients with extensive peripheral gangrene secondary to fulminant meningococcemia19. If the amputation is delayed until there is clear demarcation of both superficial and deep tissue, in many cases a Syme amputation or a through-the-knee amputation instead of a transosseous amputation can be done. These disarticulation amputations prevent stump overgrowth, provide better end-bearing surfaces, and potentially preserve distal growth plates, which is particularly important in the distal part of the femur.
In our recent experience, application of vacuum-assisted closure dressings (VAC; KCI, San Antonio, Texas) to the affected extremities of two patients with meningococcemia improved soft-tissue coverage of the residual osseous stumps and led to better success with skin grafts24.
One-fourth of our patients had some form of acquired cognitive involvement, which was mild-to-moderate in most cases. However, four children had central nervous system involvement due to brain infarction and became nonambulatory and noncommunicating and had spastic quadriplegia (GMFCS level IV or V12), requiring treatment for problems related to the static encephalopathy. Patients with mild-to-moderate developmental delay usually required only occupational and speech therapy. Fellick et al.25 found that the majority of individuals who survive purpura fulminans do not have gross neurological effects but perform worse at school than their peers. The higher prevalence of neurodevelopmental sequelae in our series may be related both to selection bias and to advances in intensive care management that have increased the survival rate of severely affected and critically ill patients with meningococcal sepsis.
In conclusion, we recommend that all patients who survive meningococcal sepsis be followed by an orthopaedic surgeon experienced in pediatric limb deformities and amputations. Early identification of possible musculoskeletal sequelae may prevent additional complications. Our data suggest that children requiring surgery for purpura fulminans are limited by their physical disability related to amputations, scarring, and abnormal bone growth. Despite advances in management, children are still at risk of developing physical and developmental limitations.