Kasabach–Merritt syndrome (KMS) is characterized by acavernous hemangioma with evidence of thrombocytopeniaand/or coagulopathy. Since Kasabach and Merritt firstdescribed this syndrome in 1940,1 approximately 300 caseshave been reported. Mortality associated with the conditionis relatively high, approaching 10–37%, because of theinvolvement of vital organs and the unresponsiveness of thecondition to treatment.2

Although the management adopted for KMS is variable,involving steroid and radiation therapy, surgery, emboliza-tion and interferon (IFN)-α, there is no cornerstone

treatment for this disease and the sporadic incidence of thedisease makes it difficult to study the treatment modalitiessystemically.

The purpose of the present study was to evaluate thevarious treatment modalities for KMS, based on a retro-spective analysis of 37 cases registering in a single centerover a period of 20 years, and to determine the most reliablemethods for treating KMS.

Methods

The present study involved 37 patients diagnosed as havingKMS who visited the Department of Pediatrics, SeoulNational University Hospital between January 1979 andJune 1999. We reviewed the age, sex, locations of thehemangiomas, clinical symptoms and hematologic data. Inaddition we analyzed patient responses to various treatmentsby examining changes in hematologic data and tumor size.

Pediatrics International (2000) 42, 620–624

Original Article

Stepwise multimodal approach in the treatment of Kasabach–Merrittsyndrome

HEE YOUNG SHIN,1 KYUNG HA RYU2 AND HYO SEOP AHN1

Departments of Pediatrics, 1Seoul National University College of Medicine and 2Ewha WomansUniversity College of Medicine, Seoul, Korea

BBaacckkggrroouunndd: The purpose of the present study was to evalutate the various treatment modalities forKasabach–Merritt syndrome (KMS) and to identify the most reliable treatment modalities.MMeetthhooddss: A retrospective analysis was performed on 37 KMS patients who were admitted to Seoul NationalUniversity Hospital between January 1979 and June 1999. Age, sex, locations of the hemangiomas, clinicalsymptoms and hematologic data were analyzed by multivariate logistical regression analysis to determine theresponse to various treatment modalities.RReessuullttss: Twenty-four of 37 patients (20 boys and 17 girls) were diagnosed during infancy. The locations ofhemangioma were superficial skin in 31 patients, combined skin and visceral organs in four patients andvisceral organs only in two patients. Tumor size was more than 5 cm in diameter in all cases, except for one.The treatment principle of KMS in our center involves a stepwise multimodal approach. Of the 37 cases,surgical resection was performed in two. Steroids were tried initially in 35 patients over a 1 week period,with an initial response rate of 11.4% (4/35). Combined steroid/radiation treatment was given in 28 cases,with a response rate of 75.0% (21/28). Four of five patients with additional interferon (IFN)-α therapyimproved. No significant correlations were found between the prognostic factors, such as age, sex, size andlocations of the hemangioma, hematologic data and the treatment modalities.CCoonncclluussiioonnss: Based on the experiences in a single center over a period of 20 years, we recommend astepwise multimodal approach for the treatment of KMS; for example, steroid, radiation therapy and IFN-αin this order. However, surgical treatment is helpful if total resection is possible.

Key words Kasabach–Merritt syndrome, single center, stepwise multimodal approach.

Correspondence: Dr Hee Young Shin, Department of Pediatrics,Seoul National University College of Medicine, 28 Yongon-dong,Chongno-gu, Seoul 110-744, Korea. Email: hyshin@plaza.snu.ac.kr

Received 4 November 1999; revised 13 March 2000; accepted18 March 2000.

Diagnostic criteria of KMS were superficial and visceralhemangioma with thrombocytopenia or abnormal hemato-logic data suggesting disseminated intravascular coagulation(DIC). Hemangioma was confirmed by gross appearancewhen it was located externally and was confirmed by eithersonography or a blood pool scan when it was locatedinternally. Biopsies were performed on two cases, but werenot performed routinely for the diagnosis of hemangioma.

The treatment principle for KMS in our center involves astepwise multimodal approach. Total resection is undertakenif it is believed possible at initial diagnosis; if not, steroidtreatment is commenced at 2 mg/kg bodyweight prednisonefor 2 weeks. The response to steroid treatment is evaluatedat the end of the second week and radiation therapy is addedfor steroid non-responders. Steroid responders are defined ashaving a complete disappearance or a significant reductionin the size of the hemangioma, a sustained platelet count ofover 150 000 /mm3 and no coagulation abnormalities. Wedefine ‘steroid non-responders’ as patients who fail to meetat least one of these criteria. Radiation therapy is given at adose of 200 cGy/day for 3 consecutive days. If there is noresponse to steroid and radiation therapy, additional doses ofradiation therapy or IFN-α are administered. Two casesadmitted recently were given IFN-α instead of radiationtherapy for fear of asymmetric bony growth. Administrationof IFN-α is started at a dose of 3 000 000 IU/m2 per day.Daily injections of IFN-α are reduced to injections onalternate days when a visible regression of the hemangiomais noted. Treatment is continued for as long as 9 months.The responses to various treatments are evaluated byexamining the decrease in the size of the hemangioma,normalization of the platelet count and satisfactory hemato-logic data.

Statistical analysis

The correlation between response to treatment and the prog-nostic factors was evaluated using multivariate logisticalregression analysis using the S P S S software package (SPSSInc., Chicago, IL, USA). P<0.05 was considered to bestatistically significant.

Results

The clinical and hematologic data, treatment and finaloutcomes are summarized in Table 1.

Patients’ ages ranged from birth to 9 years and 3 months.Twenty-four of 37 patients were diagnosed while they wereinfants. Twenty of 37 patients were boys and 17 were girls.

The location of the hemangioma was superficial skin in 35 patients, which was combined with involvement of the visceral organs in four patients. Distribution of the

superficial lesions without involvement of visceral organswere at the extremities in 11 patients, head and neck in ninepatients, the trunk in seven patients and multiple sites in fourpatients. Combined superficial and visceral hemangiomaswere found in four cases involving the abdominal wall andretroperitoneum in one case, the abdominal wall and urinarybladder in the second case, the chest, pericardium and spleenin the third case and the back, scrotum and retroperitoneumin the fourth case. Two patients had no superficial hem-angiomas; they were each found to have hemangioma of thespleen and retroperitoneum. Because these cases had nocutaneous lesions, some difficulty was experienced duringdiagnosis and the hemangiomas were found during thework-up of DIC of unknown origin and were later confirmedby ultrasonography and blood pool scans.

The tumor size measured more than 5 cm in diameter inmost cases, except for one patient. The largest tumor waslocated in the spleen (20`17 cm). The smallest heman-gioma that could have caused severe thrombocytopenia wasa hemangioma of the face (3.5`3.5 cm) and the plateletcount in this case was only 3000 /mm3.

Hematologic data included a complete blood count(CBC), prothrombin time (PT), activated partial thrombo-plastin time (aPTT), fibrinogen and fibrinogen degradationproduct (FDP). Two of four patients with combined involve-ment of internal organs were diagnosed incidentally, one asretroperitoneal hemangioma by sonography and blood poolscan during DIC work-up and the other as pericardialhemangioma during an operation for the evaluation of peri-cardial effusion. Platelet counts for patients with combinedinternal organ involvement were below 100 000 /mm3 in allcases except for one and the median platelet counts were34 000 /mm3.

Thrombocytopenia was combined with abnormalities ofPT, aPTT, fibrinogen and FDP in 34 patients, but twopatients showed only thrombocytopenia and one patientshowed abnormalities of PT, aPTT, fibrinogen and FDPwithout thrombocytopenia. Anemia (mean hemoglobin values9.6~2.9 g/dL) also developed in 28 patients.

Two patients with hemangioma in the spleen and backcompletely recovered in terms of platelet count and hemato-logic data after total resection of the hemangioma. Three of35 patients without initial operation responded to steroidtherapy within 1 week of treatment. One patient who hadhemangiomas of the abdominal wall and urinary bladder hadsurgery on the bladder hemagioma after 2 weeks of steroidtreatment with partial platelet recovery. Of the remaining 31 steroid non-responding patients, 28 were treated withradiation therapy at 600 cGy, two recently diagnosed patientswere treated with IFN-α when they did not respond tosteroids and one patient was lost to follow up after steroidtreatment only. Twenty-one of 28 patients responded toadditional radiation therapy. Of the seven patients who did

Stepwise multimodal approach 621

not respond to either steroid or radiation treatment, onepatient spontaneously recovered in terms of tumor size andlaboratory data after 1 year of follow up. Another patientdied during follow up because of a progression of DIC andtwo patients were lost during follow up. The remaining threepatients were treated with IFN-α and two responded well;one patient died due to DIC progression. The two recentlydiagnosed patients who were treated with IFN-α aftersteroid therapy are completely cured.

The median duration of follow up was 6 years and 3months and those who received steroid and radiationtreatment were followed carefully for the development ofretarded bony growth or malignancies. No radiation inducedside effects were observed during follow up.

We tried to determine the key prognostic factors byevaluating the correlation between final outcome andvarious factors, such as sex, age at diagnosis of hemangiomaor KMS, hematologic data, location, treatment methods,

622 HY Shin et al.

Table 1 Summary of clinical and hematologic data, treatment and final outcome

Patient Sex Detection time of Location Hb Platelet count Treatment Finalno. hemangioma/KMS* (g/dL) (×103 /mm3) outcome

1 M 48m/48m Spleen 4.7 4 Surgery I2 F 2m/3m Buttock, Forearm (skin) 7.9 28 Steroid, RT I3 F Birth/17m Thigh (skin) 6.2 22 Steroid, RT I4 M Birth/2m Thigh (skin) 7.7 17 Steroid, RT I5 M Birth/1d Back (skin) 21.2 55 Surgery I6 F 2m/5m Forearm (skin) 11.3 186 Steroid I7 F 2 week/1m Scalp (skin) 8.3 42 Steroid, RT I8 F Birth/5m Shoulder (skin) 7.7 40 Steroid, RT SR9 M 10d/4m Arm (skin) 11.6 26 Steroid, RT I10 M 16d/4m Neck, Calf (skin) 9.4 12 Steroid, RT I11 F 4m/5m Chest wall (skin) 8 44 Steroid, RT I12 M Birth/7m Face (skin) 8.5 51 Steroid NR13 M Birth/1m Neck (skin) 10.6 19 Steroid, RT I14 M 9m/9m Buttock, Thigh (skin) 11.6 6 Steroid, RT I15 F Birth/7m Thigh (skin) 9.8 46 Steroid, RT I16 F 1m/5m Elbow (skin) 10.5 26 Steroid, RT I17 F Birth/4m Abdominal wall (skin) 10.2 18 Steroid, RT I18 M Birth/3m Chest wall (skin) 8.7 9 Steroid, RT I19 F Birth/16d Shoulder (skin) 11.6 71 Steroid, RT D20 F 18d/2d Retroperitoneum 8.5 40 Steroid I21 F Birth/1m Arm (skin) 9.6 34 Steroid, RT I22 F 1y7m/1y11m Face, neck (skin) 8.8 53 Steroid, RT LF23 M Birth/50d Abdominal wall (skin), 6.6 13 Steroid, RT LF

retroperitoneum24 F Birth/2d Face (skin) 6.5 19 Steroid I25 M Birth/1d Abdominal wall (skin), 14.2 3 Steroid, surgery I

urinary bladder26 M Birth/2d Flank wall (skin) 13.9 59 Steroid, RT I27 F 4m/5m Flank wall (skin) 9.6 15 Steroid, RT I28 M Birth/7m Thigh (skin) 6.9 44 Steroid, RT I29 M 45d/8m Back (skin) 10.7 3 Steroid, RT I30 M 2m/6m Neck (skin) 12.2 68 Steroid, RT I31 M 9y3m/9y3m Chest wall (skin), 8.3 12 Steroid, RT, IFN-α D

pericardium, spleen32 F 1m/1m Neck (skin) 12.5 23 Steroid, RT, IFN-α I33 M 1m/1m Back, scrotum 6.7 6 Steroid, RT I

(skin), retroperitoneum34 M Birth/2m Face (skin) 8.4 21 Steroid, RT I35 M Birth/6m Groin (skin) 9.9 56 Steroid, RT, IFN-α I36 F Birth/4m Face (skin) 9.4 3 Steroid, IFN-α I37 M Birth/1m Buttock (skin) 7.5 49 Steroid, IFN-α I

*Detection of bleeding abnormalities. KMS, Kasabach–Merritt syndrome; Hb, hemoglobin; m, month; d, day; y, year; RT, radiotherapy;IFN-α, interferon-α; I, improved; SR, non-responder, but spontaneous regression later; NR, non-responder; D, death; LF, lost to follow up.

response duration and tumor size. However, no significantcorrelations were found using logistical regression models(Table 2).

Discussion

Hemangiomas are common congenital lesions in infants andchildren. They are usually benign in nature and frequentlyregress spontaneously.3 However, 3–5% of hemangiomas are complicated by their size, involvement with vital organsor concomitant coagulopathy. Kasabach and Merritt firstdescribed the association between an enlarging hemangiomawith thrombocytopenia and consumption coagulopathy in1940. Kasabach–Merritt syndrome has its highest incidencein early infancy.4 Our results concur with this observation, inthat 24 of 37 patients were diagnosed while within the infantage range.

Several factors contribute to the bleeding tendency inKMS. The possible mechanisms of consumption coagula-pathy of KMS are that platelet sequestration is merely theinitial step in a cascade of events. Within the hemangioma,there is a general decrease of blood velocity as a result of a large collection of proliferating blood vessels. As bloodpools within the lesion, contact with abnormal vascularendothelium activates an intense process of coagulation andlysis, resulting in a consumption coagulopathy.2

In spite of several treatment methods, such as surgery,steroid treatment, radiation, antifibrinolytic agents and IFN-αtreatment, KMS is still a potentially life-threateningcondition.5 In the present study, two patients died and threepatients were lost during follow up, resulting in a morbidity

of 8.1%. The causes of the deaths were progression of DIC.Hemangioma can be found in virtually any organ.

Choice of treatment modalities is empiric and no singletreatment is universally helpful. Therefore, treatment shouldbe tailored to each patient and situation. The goal oftreatment is to control consumptive coagulopathy, hemolysisand hemangioma growth.

Steroids are the first step in the treatment of KMS andother forms of invasive hemangiomas.2 The mechanisms ofsteroid therapy are unclear, but prednisone appears to increasevasoconstriction, inhibit fibrinolysis, increase platelet longevityand disrupt angiogenesis. Possible complications of steroiduse include interference with growth, reversible osteoporosisand increased susceptibility to infection. In the presentstudy, steroid therapy was shown to be effective in only11.4% of cases. When steroids are effective, the responsetypically manifests within 1 week.

Surgical intervention is an option in the treatment ofKMS.6 Because the tumor is highly vascular and there isalways the risk of concurrent DIC, surgery carries significantrisk. Total excision was performed in two patients at initialdiagnosis and in another patient after 2 weeks of steroidtreatment; no recurrences were observed.

The use of radiation in the treatment of hemangiomas isnot new and was first described by Kasabach and Merritt in1940.1 Radiation doses are variable, ranging in the literaturefrom 100 to 2400 cGy. Used alone or conjunction withsteroids, radiation has often proven beneficial, but long-termiatrogenic complications have been reported, including local failure of growth, late skin changes, destruction ofbone tissue and the potential development of secondarymalignancy.7,8 The mechanism of radiation therapy involves

Stepwise multimodal approach 623

Table 2 Correlation between final outcome and prognostic factors

Variables Prognostic factors OR 95% CI P

DependentOutcome (0=not improved, 1=improved)

IndependentSex Male (0=female) 1.37 0.14–130.40 0.89Age at diagnosis of KMS 0.86 0.64–1.16 0.33Location Including internal organ 71.09 0–0 0.63(0=cutaneous only)

Size (cm`cm) 0.99 0.96–1.03 0.97Hb (g/dL) 0.510 0.08–3.11 0.45Platelets (×103 /mm3) 1.02 0.94–1.09 0.58Treatment Steroid (0=surgery) 0 0–0 0.82Steroid+RT 0 0–0 0.86Including IFN-α 0 0–0 0.90

Note that patients who were lost during follow up were excluded from statistical analysis. Patients who regressed spontaneously wereincluded in the improved groups and patients who died during treatment or did not respond to treatment were included in the not-improvedgroup. Multivariate analysis was performed by logistic regression. P<0.05 was considered statistically significant.

OR, odds ratio; CI, confidence interval; KMS, Kasabach–Merritt syndrome; Hb, hemoglobin; RT, radiotherapy; IFN-α; interferon-α.

624 HY Shin et al.

inducing embolization within the hemangioma. In the pre-sent study, the effects of doses of 600 cGy were quiteremarkable.

In 1989, White9 reported the first case of a hemangiomatreated with IFN-α, which is used because it reportedlyinhibits the proliferation of vascular cells and has minimalside effects.9 Possible mechanisms of the action of IFN-αinclude the inhibition of the proliferation of endothelialcells, smooth muscle cells or fibroblast cells that have beenstimulated by endothelial cell or fibroblast growth factors.The side effects of IFN-α are a mild fever, mild elevationsof liver function tests and flu-like symptoms. Recently, andin spite of its high cost, IFN-α has been used as the drug offirst choice for the treatment KMS10,11 when hemangiomasare encountered in combination with thrombocytopenia andwe now adopt IFN-α as the treatment of first choice forthose patients showing little or no response to steroidtreatment.

Embolization therapy, chemotherapy, platelet aggregationand the use of fibrinolysis inhibitors are other methods forthe treatment of KMS, but their effectiveness is as yetunconfirmed. Although not common, KMS can be bothdebilitating and disfiguring and there is a risk that DIC couldoccur.

Because the patient population is small, relatively fewtrials have been conducted with a view towards providingguidance with regard to specific modes of therapy for KMS.For this reason, we have thoroughly reviewed the medicalrecords of 37 KMS patients who attended our center overthe past 20 years. We recommend a stepwise multimodalapproach in the treatment of KMS, which involves steroid,IFN-α and 600 cGy radiation therapy, in that order;

moreover, surgical treatment is helpful at anytime if totalresection is possible.

References

1 Kasabach HH, Merritt KK. Capillary hemangioma withextensive purpura: Report of a case. Am. J. Dis. Child. 1940;59: 1063–70.

2 Szlachetka DM. Kasabach–Merritt syndrome: A case review.Neonat. Network 1997; 17: 7–15.

3 Weber TR, Connors RH, Tracy TF, Bailey PV. Complex hem-angiomas of infants and children. Individualized managementin 22 cases. Arch. Surg. 1990; 125: 1017–21.

4 Esterly NB. Kasabach–Merritt syndrome in infants. J. Am.Acad. Dermatol. 1983; 8: 504–13.

5 Lopriore E, Markhorst DG. Diffuse neonatal haem-angiomatosis: New views on diagnostic criteria and prognosis.Acta Paediatr. 1999; 88: 93–7.

6 Enjolras O, Deffrennes D, Borsik M, Diner P, Laurian C.Vascular ‘tumors’ and the rules of their surgical management.Ann. Chir. Plast. Esthet. 1998; 43: 455–89.

7 Caldwell JB, Ryan MT, Benson PM, James WD. Cutaneoussarcoma arising in radiation site of a congenital hemangioma.J. Am. Acad. Dermatol. 1995; 33: 865–70.

8 Mitsuhashi N, Furuta M, Sakurai H et al. Outcome ofradiation therapy for patients with Kasabach–Merrittsyndrome. Int. J. Radiat. Oncol. Biol. Phys. 1997; 39: 467–73.

9 White CW. Treatment of hemangiomatosis with recombinantinterferon-α. Semin. Hematol. 1990; 27: 15–22.

10 Greinwald Jr JH, Burke DK, Bonthius DJ, Bauman NM,Smith RJ. An update on the treatment of hemangiomas inchildren with interferon alfa-2a. Arch. Otolaryngol. Head NeckSurg. 1999; 125: 21–7.

11 Sgonc R, Fuerhapter C, Boeck G, Swerlick R, Fritsch P, SeppN. Induction of apoptosis in human dermal microvascularendothelial cells and infantile hemangiomas by interferon-alpha. Int. Arch. Allergy Immunol. 1998; 117: 209–14.