|Home | About | Journals | Submit | Contact Us | Français|
We sought to define the characteristics that distinguish Kawasaki disease shock syndrome from hemodynamically normal Kawasaki disease.
We collected data prospectively for all patients with Kawasaki disease who were treated at a single institution during a 4-year period. We defined Kawasaki disease shock syndrome on the basis of systolic hypotension for age, a sustained decrease in systolic blood pressure from baseline of ≥20%, or clinical signs of poor perfusion. We compared clinical and laboratory features, coronary artery measurements, and responses to therapy and analyzed indices of ventricular systolic and diastolic function during acute and convalescent Kawasaki disease.
Of 187 consecutive patients with Kawasaki disease, 13 (7%) met the definition for Kawasaki disease shock syndrome. All received fluid resuscitation, and 7 (54%) required vasoactive infusions. Compared with patients without shock, patients with Kawasaki disease shock syndrome were more often female and had larger proportions of bands, higher C-reactive protein concentrations, and lower hemoglobin concentrations and platelet counts. Evidence of consumptive coagulopathy was common in the Kawasaki disease shock syndrome group. Patients with Kawasaki disease shock syndrome more often had impaired left ventricular systolic function (ejection fraction of <54%: 4 of 13 patients [31%] vs 2 of 86 patients [4%]), mitral regurgitation (5 of 13 patients [39%] vs 2 of 83 patients [2%]), coronary artery abnormalities (8 of 13 patients [62%] vs 20 of 86 patients [23%]), and intravenous immunoglobulin resistance (6 of 13 patients [46%] vs 32 of 174 patients [18%]). Impairment of ventricular relaxation and compliance persisted among patients with Kawasaki disease shock syndrome after the resolution of other hemodynamic disturbances.
Kawasaki disease shock syndrome is associated with more-severe laboratory markers of inflammation and greater risk of coronary artery abnormalities, mitral regurgitation, and prolonged myocardial dysfunction. These patients may be resistant to immunoglobulin therapy and require additional antiinflammatory treatment.
Although the cardiac complications of KD are well known, hemodynamic instability is unusual in the acute phase of illness, except as a complication of intravenous IVIG administration.
We have observed shock and hypotension with increasing frequency in newly diagnosed KD. Compared with hemodynamically normal KD, KDSS is associated with increased inflammation, platelet consumption, IVIG resistance, coronary artery abnormalities, mitral regurgitation, and myocardial dysfunction.
Kawasaki disease (KD) is the most common cause of acquired heart disease in the pediatric age group and results in permanent damage to the coronary arteries in up to 25% of untreated children.1 However, hemodynamic instability during the acute phase of the illness, not related to the infusion of intravenous immunoglobulin (IVIG), is uncommon. In 2006–2007, we treated an increased number of children with KD who required hemodynamic support in the critical care setting. This prompted us to define this complication of KD and to compare the clinical characteristics and outcomes of patients with KD shock syndrome (KDSS) with those of patients with hemodynamically normal KD.
Since January 2001, the KD Research Center at Rady Children’s Hospital San Diego has maintained a prospective database on patients treated for KD. From this database, we identified all patients with acute KD who had been admitted to this freestanding children’s hospital, which serves as the principal pediatric inpatient facility for a population base of ~3 million. We included all patients treated during the 4-year period from May 2003 through April 2007 who had fever for ≥3 days but ≤10 days and who fulfilled ≥4 of 5 diagnostic criteria (rash, conjunctival injection, cervical lymphadenopathy, changes in the oral mucosa, and changes in the extremities) or 3 criteria plus coronary artery abnormalities documented through echocardiography.2 Patients with KD were divided into a group with shock (KDSS group) and a group in hemodynamically normal condition before infusion of IVIG. We defined hypotension and shock according to previously published guidelines, and we made the diagnosis of KDSS if the sustained presence of any of the following conditions caused the treating clinicians to initiate volume expansion, infusion of vasoactive agents, or transfer to an intensive care setting: systolic hypotension for age (infants 0–28 days of age, <60 mm Hg; infants 1–12 months of age, <70 mm Hg; children 1–10 years of age, <70 + [2 × age] mm Hg; youths >10 years of age, ≤90 mm Hg3,4), a decrease in systolic blood pressure from baseline of ≥20%, or clinical signs of poor perfusion (tachycardia, prolonged capillary filling time, cool extremities, diminished pulses, oliguria, or mental status changes not accounted for by other conditions such as fever or ambient temperature) regardless of measured blood pressure.3,5 Patients with KD who were admitted during the same period and did not meet the criteria for the definition of shock formed the comparison group with hemodynamically normal KD.
Although no patients with KD presented with shock in the prospective database before 2003, we postulated that earlier cases might have been misdiagnosed and therefore never included in our database. To identify patients with KD who were treated for shock but were misclassified with other diagnoses, we conducted a computerized search, on the basis of International Classification of Diseases, Ninth Revision codes, for patients discharged from the critical care unit between January 1, 2001, and August 31, 2006, with diagnoses of toxic shock syndrome (n = 6), systemic inflammatory response syndrome with (n = 160) or without (n = 109) organ dysfunction, septicemia (n = 28), shock (unspecified) (n = 66), or viremia (n = 4). We reviewed the patients’ hospital records to identify clinical features of KD.
For all patients with KD, we accessed the following prospectively collected data: demographic data (age, gender, ethnicity, and number of days of fever at presentation); laboratory values obtained before the administration of IVIG (white blood cell and differential counts, platelet count, erythrocyte sedimentation rate, and concentrations of hemoglobin, C-reactive protein, alanine transaminase, and γ-glutamyl transpeptidase); laboratory evaluations of coagulation function within 72 hours before or after IVIG administration; response to IVIG; and echocardiographic data. We converted neutrophil and band counts reported as percentages to absolute neutrophil and band counts and hemoglobin levels to age-adjusted z scores.6 We defined IVIG resistance as persistent or recrudescent fever (temperature of ≥38.0°C, measured rectally or orally) ≥36 hours after completion of the IVIG infusion (2 g/kg).2
We compared echocardiographic measurements between the KDSS group and the subset of the hemodynamically normal KD group for which complete echocardiographic data were available (n = 86). Echocardiography (Sequoia; Siemens Medical Systems, Malvern, PA) was performed with standard techniques during the acute (echocardiogram 1; 0–10 days after disease onset), subacute (echocardiogram 2; 11–21 days), convalescent (echocardiogram 3; 22–90 days), and chronic (echocardiogram 4; 91–365 days) phases. We measured the internal diameters of the proximal right coronary artery and left anterior descending coronary artery, as well as the aortic root at 4 locations (annulus, sinus, sinotubular junction, and ascending aorta). The coronary arteries were classified as normal (luminal internal diameters of the right coronary artery and left anterior descending coronary artery measured with echocardiography were <2.5 SD units from the mean normalized for body surface area; z score of <2.5), dilated (2.5 ≤ z score < 4.0), or aneurysmal (saccular or fusiform dilation of a coronary artery segment with a z score of ≥4.0).7 We defined the lower limit of the reference range for the ejection fraction as 54%, and we defined abnormal valvular insufficiency as retrograde flow of ≥1+.8
Parameters of ventricular diastolic function included mitral inflow velocities during early diastolic filling (E wave velocity) and atrial contraction (A wave velocity), deceleration time (time, in milliseconds, from the peak of the E wave to baseline), and Doppler measurement of tissue velocity at the lateral mitral annulus during early diastolic filling (E’ velocity). Data calculated from the diastolic measurements included the mitral E wave velocity/A wave velocity ratio and the E wave velocity/E’ velocity ratio. Diastolic parameters were compared with published normal values9-11 and, when regression equations were available, z scores were generated.
We performed bivariate comparisons by using the Mann-Whitney U test for medians and the χ2 test or Fisher’s exact test for dichotomous or ordinal variables, as appropriate. For laboratory values that were outside the dynamic range of the test (eg, erythrocyte sedimentation rate of >140 mm/hour), the maximal or minimal value detected in the assay was used, as appropriate. We performed statistical analyses with SPSS 11.5 for Windows (SPSS, Chicago, IL).
Each of the 373 patients discharged from the critical care unit with shock syndromes between January 2001 and August 2006 either had an unequivocal alternate diagnosis with supporting test results or did not meet the diagnostic criteria for KD. Therefore, we identified no reasonably likely cases of missed KDSS. Between May 2003 and April 2007, 13 patients with KDSS among 187 patients with KD were treated in a critical care unit (n = 12) or closely monitored setting (n = 1) for shock (Table 1 and Fig 1). All received volume resuscitation (median: 53 mL/kg) and 7 (54%) received infusions of vasoactive agents for 3 to 135 hours (median: 68 hours). The following agents were used singly or in combination: dobutamine, dopamine, epinephrine, milrinone, and norepinephrine. Compared with patients with hemodynamically normal KD, patients with KDSS were more often female and had greater proportions of bands, lower platelet counts, lower age-adjusted hemoglobin z scores, and higher C-reactive protein concentrations (Table 2). Age, ethnicity, illness duration, and other laboratory characteristics did not differ between the groups. Although the difference was not statistically significant, patients reporting Hispanic ethnicity were nearly twice as common in the KDSS group and Asian patients were not represented. Of the 11 patients who underwent a coagulation evaluation, 5 patients (45%) had evidence of consumptive coagulopathy, with low platelet counts (<150 × 109 cells per L), positive D-dimer results, and prolonged partial thromboplastin times for age (Table 1). The remaining 6 all demonstrated ≥1 abnormality in their coagulation evaluations. IVIG resistance was more common in KDSS, affecting 6 patients (46%) (Table 2), who required additional IVIG or infliximab to control their inflammation.
The KDSS group had more-severe abnormalities in numerous echocardiographic measures of coronary artery anatomic features and ventricular function (Table 3). The maximal left anterior descending coronary artery and right coronary artery z scores were higher in the KDSS group, but aortic root z scores were not. Abnormally low ejection fractions and mitral regurgitation occurred more frequently in the KDSS group; however, both parameters were normal by the subacute phase (echocardiogram 2). None of the echocardiograms performed in the acute phase of KDSS revealed aortic regurgitation or pericardial effusions. All patients with KDSS underwent echocardiography during the subacute period, and 6 had convalescent echocardiograms. None had residual mitral regurgitation, and coronary artery diameters were normal in 5 of the 6 convalescent studies performed to monitor acute dilation. None of the electrocardiograms or echocardiograms obtained during the acute phase or during the follow-up period revealed evidence of ischemia or infarction.
Striking abnormalities of diastolic function were noted in the KDSS group (Table 4). Fusion of E and A waves in acute-phase echocardiograms limited the assessment of diastolic function for all except 2 patients in the KDSS group. Mitral E wave velocities were abnormal on echocardiogram 1 for 5 of 13 patients with KDSS and continued to be abnormally low on echocardiograms 3 and 4 for 3 of 5 patients. On echocardiogram 2, the mitral E wave velocity/A wave velocity ratio was more frequently decreased in KDSS, which suggests impaired ventricular relaxation. Abnormally low E wave velocity/A wave velocity ratios persisted through the convalescent and chronic studies for patients with KDSS. At echocardiogram 2, elevated mitral E wave velocity/E’ velocity ratios for 6 of 7 patients with KDSS suggested persistent elevation of left ventricular end diastolic pressures, and shortened deceleration times for 3 of 12 patients suggested decreased ventricular compliance. E wave velocity/E’ velocity ratios and deceleration times were not available for sufficient numbers of patients for comparison with data for patients without shock in later stages, but deceleration times were shortened for 1 of 3 patients and 2 of 5 patients with KDSS in the convalescent and chronic stages, respectively.
We report the occurrence of shock and hypotension unrelated to IVIG administration in 13 of 187 patients with KD over a 4-year period, a phenomenon that we had not observed previously at our center. Compared with patients with hemodynamically normal KD, patients with KDSS were more likely to be female, to have laboratory findings consistent with greater inflammation, and to have impaired systolic and diastolic function. Patients with KDSS also exhibited resistance to IVIG more often and had higher rates of coronary artery dilation and aneurysm formation.
Although the most-overt cardiovascular disturbances resolved promptly with therapy, evidence of abnormal ventricular diastolic function persisted into the chronic phase of follow-up monitoring. Ventricular relaxation, which is an active process in early diastole, allows normal E wave velocity. Early impairment in relaxation results in a depressed E wave velocity/A wave velocity ratio, as seen in our KDSS population. Decreases in ventricular compliance and elevated filling pressures result in shorted deceleration times and elevated E wave velocity/E’ velocity ratios, both of which were present in the subacute phase of KDSS. The cause of persistent ventricular diastolic dysfunction and its long-term significance remain unknown. However, there may be long-term sequelae of KDSS even after coronary artery abnormalities have resolved. Lack of complete diastolic data for the group with hemodynamically normal KD limited our comparisons.
The cause and factors contributing to the development of KDSS are unknown. Although the number of days of fever at diagnosis did not differ between patients with and without KDSS, laboratory test results before treatment and disease outcomes suggested greater underlying inflammation. The inciting factors are unknown, but the high frequency of low-grade consumptive coagulopathy suggests more-intense vasculitis in KDSS. Coagulopathy and platelet activation have been described in acute KD, although the process usually is self-limited and does not require specific therapy beyond treatment of the underlying inflammation.12
The prevalence of KDSS was highest in 2006 and 2007 (Fig 1), and we found no cases consistent with KDSS before 2003, through a retrospective review of other diagnostic categories that included shock, which suggests that KDSS is an increasingly recognized phenomenon in our community. Previous reports described patients with KD without evidence of coronary artery occlusion who experienced shock and hypotension in association with acute respiratory distress syndrome,13 aortic valvular damage and severe aortic regurgitation,14 renal failure and relative volume overload,15 and severe renal failure and myocarditis.16 Our study, which emphasizes the temporal clustering of cases of shock in KD, and the study of a similar series of patients with KDSS from Denver, Colorado,17 should alert clinicians to the possible increasing prevalence of this complication of acute KD.
We recognize several limitations to this study. In the absence of a previously accepted definition for KDSS, our inclusion criteria combined hypotension, relative blood pressure changes, and clinical measures of hypoperfusion. Normative values for pediatric hypotension are less well established than those for hypertension. We used lower-limit values for systolic hypotension predicted by a simple, age-based formula.3,4 Comparable lower-limit values are recommended by other sources, including a consensus definition used for pediatric sepsis18,19 and the extrapolation of 5th percentile values of systolic blood pressure from better-established 50th and 95th percentile values.20 We incorporated a relative decrease in blood pressure into our definition of KDSS, which is not part of the traditional definition of shock. However, our criterion (≥20% decrease from baseline values) is consistent with the magnitude of changes deemed clinically significant by other pediatric investigators, including relative decreases of 10% to 30% and absolute decreases of 20 to 30 mm Hg in the context of pediatric sedation,21,22 anesthetic induction,23-25 emergency airway management,26 and anaphylaxis.27 Our inclusion of purely clinical criteria for shock might allow patients with mild, potentially self-limited, hemodynamic disturbances to meet our definition of KDSS. However, our definition is in keeping with those of the Pediatric Advanced Life Support3 and Advanced Pediatric Life Support5 courses, which emphasize early detection and treatment of shock without a requirement for frank hypotension or laboratory or invasive measures of cardiac output and tissue perfusion.
Despite prospective entry of data into the KD database, the recent recognition of KDSS was unanticipated, and patient characteristics, management practices, and other regional and global phenomena contributing to or predicting KDSS and treatment failure might have been overlooked. Similarly, it is conceivable that brief episodes of hemodynamic instability escaped clinical detection either in the KD population before 2003 or in our comparison group of patients without shock after 2003. However, there have been no changes in the referral patterns to our center, which is the principal pediatric tertiary care center for a 3-county region, or in the maintenance or content of the KD database. Nursing care and documentation standards for patients undergoing evaluation and treatment for KD also have remained standardized. Finally, the retrospective nature of our search for previously misclassified patients with KDSS and our small sample of patients with KDSS limit our ability to describe the full spectrum of this phenomenon and to identify independent risk factors through multivariate analyses.
We have observed shock and hypotension requiring critical care support with increasing frequency among patients with acute KD. These patients seem to have greater incidences of coronary artery abnormalities and IVIG resistance. Additional research is required to elucidate the pathogenesis of this phenomenon, to identify risk factors and predictors for the development of KDSS, and to characterize more fully the long-term abnormalities of diastolic dysfunction. However, clinicians who treat pediatric patients with shock and those who treat patients with acute KD should maintain a heightened awareness of KDSS and should be prepared to institute hemodynamic support and to administer repeat doses of IVIG or adjunctive antiinflammatory therapies. Awareness of KDSS is particularly important for emergency department personnel, who are likely to represent the first point of medical contact for these very ill patients, and for critical care personnel, who rarely care for patients with KD.
This work was supported in part by grants from the National Heart, Lung, and Blood Institute, National Institutes of Health (grant RO1 HL69413 to Drs Burns and Kanegaye and grant K24-HL074864 to Dr Burns).
The authors have indicated they have no financial relationships relevant to this article to disclose.