An explorative study on quality of life and psychological and cognitive function in pediatric survivors of septic shock

Madelon B. Bronner, MA;1 Hendrika Knoester, MD, PhD;1 Jeanine J. Sol, MD; Albert P. Bos, MD, PhD; Hugo S. A. Heymans, MD, PhD; Martha A. Grootenhuis, MSc, PhD

Objective: To evaluate self-reported health-related quality of scores were comparable to the age-related Dutch norm popula- life, anxiety, depression, and cognitive function in pediatric septic tion. Depression scores were significantly better than the norm shock survivors. population, whereas cognitive function was significantly lower Design: A retrospective cohort study. than the norm population. We found that 44% of the children had Setting: A 14-bed tertiary pediatric intensive care unit. cognitive scores <25% of the norm population. Young age at the Patients: Children aged >8 yrs at the time of the follow-up time of pediatric intensive care unit admission was predictive of who were admitted between 1995 and 2004 for septic shock. cognitive problems, and cognitive problems were associated with Inotropic and or vasoconstrictive agents were administered to lower emotional function. these patients for >24 hrs. Conclusions: In this group of septic shock survivors, health- Intervention: Health-related quality of life was assessed with related quality of life, anxiety, and depression are equal to or the KIDSCREEN-52, anxiety with the State Trait Anxiety Inventory slightly better than the age-related Dutch norm population. Cog- for Children, depression with the Children’s Depression Inventory, nitive function is decreased, especially in children admitted at and cognitive function with the cognitive scale of the TNO-AZL younger ages. Follow-up studies with adequate neuropsycholog- Children’s Quality of Life Questionnaire Child Form. ical testing are warranted to evaluate the association between Measurements and Main Results: Fifty of 82 eligible pediatric septic shock, cognitive function, and risk factors for cognitive septic shock survivors were evaluated. The median age of the problems. (Pediatr Crit Care Med 2009; 10:636–642) children at pediatric intensive care unit admission was 4.2 yrs KEY WORDS: outcome assessment (health care); follow-up stud- (range, 0.0–17.0 yrs); the median age at follow-up was 10.7 yrs ies; health status; quality of life; shock septic; anxiety; depression (range, 8.0–20.4 yrs). Health-related quality of life and anxiety

he incidence and mortality quelae has resulted in “patient-centered chronic medical condition. It is a self- rate of sepsis in children are outcome” research in adult intensive care report questionnaire that evaluates lower compared with those in unit (ICU) survivors (4). Studies on long- HRQoL in children aged 8 to 18 yrs (7). adults, but sepsis is one of the term sequelae and HRQoL in pediatric The self-report questionnaire is based on Tleading causes of death in children (1). sepsis survivors are scarce. psychosocial aspects of well-being rather Adult sepsis survivors suffer from persist- Quality of life is defined as an individ- than physical function. ing symptoms, such as dyspnea, fatigue, ual’s perception of his or her position in In the few follow-up studies evaluating depression, impaired cognitive function life in the context of culture and value HRQoL in pediatric ICU (PICU) survivors, and functional status, and reduced systems as well as in relationship to his or HRQoL was satisfactory in the majority of health-related quality of life (HRQoL) her goals, expectations, standards, and children (8–12). In studies evaluating compared with the general population (2, concerns. HRQoL is defined as quality of HRQoL in pediatric septic shock survi- 3). Recognition of these long-term se- life in which a dimension of personal vors, HRQoL was decreased on physical judgment of one’s health and disease is domains (13, 14). Long-term psychologi- added (5). In the case of children, HRQoL cal problems like posttraumatic stress 1Both authors are first author. From the Psychosocial Department (MBB, MAG), is also influenced by factors, such as the disorder have been reported, but other Pediatric Intensive Care Unit (HK, JJS, APB), and the ability to participate in peer groups and psychological problems, such as anxiety Department of Pediatrics (HSAH), Emma Children’s the ability to keep up with developmental and depression, are hardly studied in pe- Hospital, Academic Medical Center, Amsterdam, Neth- activities. Difficulties in measuring diatric ICU survivors (15–18). Evaluation erlands. HRQoL in children include a lack of con- of cognitive function in pediatric menin- This study was supported, in part, by the Aca- demic Medical Center, Amsterdam, Netherlands. sensus on suitable (cross-cultural) in- gitis survivors and adult septic shock sur- The authors have not disclosed any potential con- struments and the need for different in- vivors does show substantial problems flicts of interest. struments in different age groups (6). that interfere with daily life (19–23). Pos- For information regarding this article, E-mail: Recently, the KIDSCREEN-52 was devel- sible pathophysiologic mechanisms me- [email protected] Copyright © 2009 by the Society of Critical Care oped in Europe as a generic, cross- diating cognitive dysfunction in septic Medicine and the World Federation of Pediatric Inten- national questionnaire that evaluates shock survivors include hypoxemia, sed- sive and Critical Care Societies HRQoL regardless of whether children atives or analgesics, hypotension, delir- DOI: 10.1097/PCC.0b013e3181ae5c1a are in good health or suffer from a ium, hyperglycemia, and sepsis and in-

636 Pediatr Crit Care Med 2009 Vol. 10, No. 6 flammation (20). Based on these uses questions derived from focus groups of Cognitive function was evaluated by the observations, we believe that follow-up children and adolescents across Europe. It is cognitive scale of the TNO-AZL Children’s research on pediatric septic shock survi- applicable for both healthy and chronically ill Quality of Life Questionnaire Child Form that vors and their families is needed to eval- children and adolescents aged between 8 and included eight items (31, 32). These items uate long-term sequelae and to improve 18 yrs. The KIDSCREEN-52 consists of three measure limitations concerning cognitive support after discharge. sociodemographic questions (gender, age, and functioning and school performances (child’s The aim of our study was (1) to eval- educational level) and ten domains of HRQoL: ability to pay attention, understand school- uate HRQoL, psychological function Physical well-being, psychological well-being, work, understand what others say, perform (anxiety and depression), cognitive func- moods and emotions, self-perception, auton- arithmetic, read, write, learn, and say what tion, and educational level in children omy, parent relations and home life, financial he/she means). This self-report questionnaire is a generic Dutch instrument that measures who survived septic shock compared with resources, peers and social support, school health status problems and is weighted by the normative data from the Dutch popula- environment, and bullying. Items were scored on a 5-point scale. Within each domain, item impact of health problems on the child’s well- tion; (2) to explore possible medical de- being. If a problem occurred in the last few terminants that are related to cognitive scores were summed and transformed to Ra- sch person parameters by an algorithm that weeks, the child can indicate how he or she problems; and (3) to determine whether gives children in the reference population a felt about this problem on a 4-point Likert cognitive function is related to HRQoL mean score of 50 with a standard deviation of scale: (very) good (3); not so well (2); rather and psychological function. 10 (26). This instrument was validated in a bad (1); and bad (0). The score was calculated Dutch population of 1960 children, with an by adding up item scores, and higher scores MATERIALS AND METHODS age distribution of 8 to 11 yrs (n ϭ 641) and indicated a better quality of life. Maximum 12 to 18 yrs (n ϭ 1270). The normative groups domain scores for the cognitive scale were 32. ␣ Protocol and Procedure for our analysis were based on these data. The The internal consistency (Cronbach’s ) was internal consistency (Cronbach’s ␣) was satis- satisfactory (0.79). This instrument was vali- This study was part of a retrospective study factory and ranged from 0.68 to 0.88 in all dated in a Dutch population of 2331 children, ϭ in which we evaluated cardiac function, scales. with an age distribution of 8 to 11 yrs (n ϭ HRQoL, psychological function (anxiety and Psychological function measures anxiety 1078) and 12 to 15 yrs (n 1253). The nor- depression), and cognitive function in pediat- and depression. Anxiety was evaluated by the mative groups for our analysis were based on these data. ric septic shock survivors. The PICU of the State-Trait Anxiety Inventory for Children Besides a scale score, the percentage of Emma Children’s Hospital/Academic Medical (27). The trait scale, a 20-item self-report Center Amsterdam is a tertiary pediatric ICU children with cognitive problems was deter- scale, measures differences between children with 14 beds that admits medical, surgical, mined. The definition of children with cogni- in their tendency to experience anxiety states. and trauma patients from the greater Amster- tive problems was based on the value of the The scale ranges from 20 to 60 points, and dam area. Previously, healthy children who 25th percentile in the norm population (33). children with high scores are more prone to survived septic shock in our PICU between According to this concept, an individual scor- respond with anxiety to situations perceived as 1995 and 2004 were included in this study. ing below the 25th percentile norm is in the threatening than children with low scores. The term previously healthy was defined as quarter of the population most impaired. The This instrument was validated in a Dutch pop- having no need for medical supervision at any 25th percentile for healthy children (8 to 11 ulation of 1229 children, with an age distribu- time before PICU admission. Inclusion criteria yrs of age) is a score of 27, and the score for tion of 8 to 12 yrs (n ϭ 643) and 13 to 17 yrs were survival of the clinical diagnosis of septic healthy adolescents (12 to 15 yrs of age) is 26. (n ϭ 586). Boys (n ϭ 596) and girls (n ϭ 633) shock according to the Conference Consensus Patient characteristics (age at PICU admis- differed significantly on State Trait Anxiety Criteria (24, 25), the administration of inotro- sion; length of PICU stay; [length of] artificial Inventory for Children in the Dutch popula- pic and/or vasoconstrictive agents for Ն24 ventilation; risk of mortality; highest creati- tion. The normative groups for our analysis hrs, and age Ն8 yrs at the time of the fol- nine and lactate measurements during PICU low-up study. Children were invited to visit were based on these data (28). The internal stay; causative organisms; and number, dos- ␣ our outpatient follow-up clinic where they consistency (Cronbach’s ) was satisfactory ages, and duration of vasoactive agents) were were evaluated for physical sequelae, HRQoL, and ranged from 0.73 to 0.83. obtained retrospectively from medical records anxiety, depression, and cognitive function. Depression was evaluated by the Children’s and the Patient Data Management System. In Physical sequelae were defined as any physical Depression Inventory (29). This questionnaire our unit, it is no standard procedure to per- complaints or abnormalities found at the out- contains 27 items, each of which consists of form a lumbar puncture in septic shock pa- patient follow-up clinic by physical examina- three statements. For each item, the individ- tients; therefore, meningitis could not be eval- tion (e.g., neurologic abnormalities). Children ual was asked to select the statement that best uated as a risk factor for decreased HRQoL, with language barriers were excluded due to describes his or her feelings for the last 2 and psychological and cognitive function. the inability to complete Dutch question- weeks. The Children’s Depression Inventory is naires. In addition, children Ͻ8 yrs were ex- designed to provide information about the Statistical Analysis cluded because they were unable to complete presence and severity of depressive symptoms. the questionnaires by themselves. Written in- This instrument was validated in a Dutch pop- The Statistical Package for Social Sciences formed consent was obtained from all partic- ulation of 886 children, with an age distribu- (SPSS), Windows version 12.0, was used for all ipating parents and children. The Medical Eth- tion of 8 to 12 yrs (n ϭ 673) and 13 to 17 yrs analyses. First, missing values were handled ical Board of the Academic Medical Centre (n ϭ 213). Boys (n ϭ 419) and girls (n ϭ 467) according to the guidelines given in the man- Amsterdam has approved the study protocol. differed significantly on Children’s Depression uals for the relevant questionnaires. In gen- Inventory in the Dutch population. The nor- eral, data were imputed if children completed Questionnaires mative groups for our analysis were based on at least 90% of the questionnaire by mean these data (30). The internal consistency scores of the other items. Two missing items HRQoL was evaluated by the KIDSCREEN- (Cronbach’s ␣) was satisfactory and ranged were allowed for the cognitive function scale. 52, a generic self-report questionnaire that from 0.71 to 0.89. Second, Mann-Whitney U tests and ␹2 tests

Pediatr Crit Care Med 2009 Vol. 10, No. 6 637 were used to compare participants and non- Table 1. HRQoL of survivors of septic shock compared with Dutch norm data participants. Third, differences in HRQoL, anxiety, depression and cognitive function be- Child Norm Adolescent Norm ϭ ϭ ϭ ϭ tween patients and the norm group were an- (n 31) (n 641) (n 19) (n 1270) Ϯ Ϯ Ϯ Ϯ alyzed by one-sample Student’s t tests. In ad- M SD Effect Size M SD M SD Effect Size M SD dition to this, effect sizes were calculated by Physical well-being 59.2 Ϯ 11.2 0.2 57.3 Ϯ 9.5 52.5 Ϯ 13.8 0.1 50.4 Ϯ 9.4 dividing the difference in mean scores be- Psychological well-being 60.5 Ϯ 9.1 0.5 55.7 Ϯ 9.1 57.3 Ϯ 9.2 0.6 51.9 Ϯ 8.4 tween the patients and the norm group by the Moods and emotions 56.4 Ϯ 11.1 0.4 52.5 Ϯ 9.5 52.5 Ϯ 9.1 0.2 50.5 Ϯ 9.6 standard deviation of the scores from the Self-perception 61.6 Ϯ 9.8 0.4 57.4 Ϯ 9.8 51.0 Ϯ 9.4 0.1 49.9 Ϯ 9.3 norm group. According to Cohen, effect sizes Autonomy 56.2 Ϯ 9.5 0.0 56.5 Ϯ 8.6 55.7 Ϯ 7.3 0.2 53.5 Ϯ 9.1 of about 0.2 were considered to be small, effect Parent relations and 58.4 Ϯ 8.6 0.3 55.5 Ϯ 8.4 54.8 Ϯ 6.7 0.3 52.3 Ϯ 9.3 sizes of about 0.5 to 0.8 were considered to be home life Ϯ Ϯ Ϯ Ϯ moderate, and effect sizes of Ն0.8 were con- Financial resources 51.9 10.7 0.1 51.4 10.6 56.1 7.0 0.4 52.7 9.2 Peers and social support 55.0 Ϯ 10.8 0.2 53.1 Ϯ 9.1 55.0 Ϯ 7.1 0.4 51.7 Ϯ 9.4 sidered to be large (34). Fourth, cognitive School environment 59.0 Ϯ 11.1 0.1 58.4 Ϯ 10.2 53.0 Ϯ 9.7 0.3 50.7 Ϯ 8.7 function was also evaluated by analyzing dif- Bullying 48.0 Ϯ 11.4 0.0 48.1 Ϯ 11.0 47.0 Ϯ 9.0 0.2 48.8 Ϯ 10.1 ferences in educational levels between patients and healthy control subjects (data from Dutch HRQoL, health-related quality of life; Mean Ϯ SD, mean Ϯ standard deviation. Health Statistics); for this analysis, we used a p Ͻ .005 study group vs. norm data. Higher scores represent better HRQOL. one-sample Student’s t test. Fifth, cognitive function scores were dichotomized to deter- dren completed the questionnaires. No found in all depression scores, except in mine cognitive problems. A score Ͻ25th per- statistically significant differences were boys attending secondary school (Ն13 yrs centile value for the appropriate age and gen- found between patient characteristics of age). Children surviving septic shock der population stratum indicated perceived impaired cognitive function. To analyze (age of child at PICU admission and fol- reported less depression compared with whether children with and without cognitive low-up study, gender, length of stay, the Dutch norm population. Moderate problems differed with regard to medical char- length of artificial ventilation, risk of and large effect sizes were found for all acteristics, Mann-Whitney U tests and ␹2 tests mortality, highest creatinine and lactate, anxiety and depression scores, except for were used. Multivariate linear regression anal- and physical sequelae) of participants anxiety scores in boys attending second- ysis with stepwise backward variable elimina- (n ϭ 50) and nonparticipants (n ϭ 32) ary school (Table 2). tion was performed to explore risk factors (data not shown). Mean follow-up time (gender, length of stay in PICU, length of was 6.5 yrs (range, 1.5–10.1 yrs). No sta- artificial ventilation, risk of mortality, the age tistically significant differences were Cognitive Function of child at the time of follow-up study, and the found between participants and nonpar- Fifty children completed the cognitive age of child at PICU admission) for cognitive ticipants with regard to causative organ- scale of the TNO-AZL Children’s Quality function. Finally, the relationship among cog- isms and number, duration and dosages nitive function, HRQoL, anxiety, and depres- of Life Questionnaire Child Form. How- of vasoactive agents (data not shown). Յ sion was analyzed by calculating Spearman ever, only 45 children ( 15 yrs) were compared with norm data. Children had rank correlations. HRQoL Bonferroni correction was carried out to significantly worse cognitive function compensate for multiple testing. A signifi- Fifty children completed the KID- compared with the norm population. Ad- cance level of p Ͻ .005 was used in all tests, SCREEN-52. No statistically significant dif- olescents reported scores comparable except for comparisons of patient characteris- ferences were found in HRQoL in the 31 with the norm population (Table 2). tics. In these tests (participants vs. nonpartic- children and the 19 adolescents compared Twenty (44%) of the 45 children scored Ͻ th ipants), p Ͻ .05 was used as the significance with age-related Dutch norm populations. 25 percentile of the norm population, level. Moderate effect sizes were found in the indicating cognitive problems. psychological well-being domain in both Education. Thirty-seven participants ϭ RESULTS age groups; the study group was shown to (n 15 girls, 22 boys) attended primary have better psychological well-being than school at the time of the follow-up study, ϭ Participants the norm group. Effect sizes for the other whereas 13 (n 8 girls, 5 boys) attended domains were small (Table 1). secondary school. Five (14%) of the pri- From 1995 through 2004, 124 patients mary school students and two (15%) of survived admission to our PICU with sep- Psychological Function the secondary school students visited tic shock and inotropic and/or vasocon- schools for special education. In total, strictive support for Ն24 hrs. Of these Fifty children (n ϭ 23 girls, 27 boys) seven (14%) of 50 children attended spe- 124 patients, 82 children were aged Ն8 completed the anxiety questionnaire and cial education schools and five (71.4%) of yrs at follow-up and eligible for this 48 children (n ϭ 22 girls, 26 boys) com- these seven children had cognitive prob- study. Seventeen of these patients were pleted the depression questionnaire. lems (Ͻ25th percentile). In the Dutch lost to follow-up. Sixty-five children were Overall good psychological function was population, generally 3% of children and invited to participate in the study. Eight determined. No statistically significant 3.5% of adolescents attend special educa- patients refused participation (for geo- differences were found in the anxiety tion schools (data from Dutch Health graphical reasons or lack of interest). scores, except in girls surviving septic Statistics, www.cbs.nl). The number of Seven patients’ data were missing due to shock. They reported less anxiety com- children attending special education at language problems or incomplete ques- pared with the Dutch norm population. primary school (t ϭ 1.845, df ϭ 36, p ϭ tionnaires. Fifty (77%) of these 65 chil- Statistically significant differences were .073) and adolescents at secondary school

638 Pediatr Crit Care Med 2009 Vol. 10, No. 6 Table 2. Anxiety, depression, and cognitive function compared with Dutch norm data Table 4. Correlations between cognitive function and HRQoL domains, anxiety, and depression All Children With Septic Norm n Shock M Ϯ SD Effect Size Mean Ϯ SD Cognitive Function

Anxietya n ϭ 45 Correlations p Girls 23 28.3 Ϯ 6.7e 0.7 33.0 Ϯ 6.5 Primary school 15 29.1 Ϯ 7.2 0.6 33.4 Ϯ 6.5 HRQoL Secondary school 8 26.9 Ϯ 5.8 0.9 32.5 Ϯ 6.7 Physical well-being 0.351 .02 Boys 27 28.2 Ϯ 6.1 0.3 30.0 Ϯ 6.0 Psychological well-being 0.375 .01 Primary school 22 27.6 Ϯ 6.3 0.7 31.2 Ϯ 5.5 Moods and emotion 0.568a .00 Secondary school 5 31.0 Ϯ 4.1 0.4 28.7 Ϯ 6.4 Self-perception 0.335 .02 Depressionb,c Autonomy 0.363 .02 Girls 22 5.1 Ϯ 4.1e 0.6 9.3 Ϯ 6.5 Parents relation and home 0.483a .00 e Primary school 15 5.4 Ϯ 4.4 0.5 8.9 Ϯ 6.5 life e Secondary school 7 4.5 Ϯ 3.5 1.0 10.5 Ϯ 6.5 Financial resources 0.274 .09 e Boys 26 5.5 Ϯ 4.1 0.5 8.2 Ϯ 5.7 Peers and social support 0.396 .01 e Primary school 21 5.5 Ϯ 4.5 0.5 8.4 Ϯ 5.9 School environment 0.367 .02 Secondary school 5 5.4 Ϯ 2.1 0.5 7.7 Ϯ 4.9 Bullying 0.610a .00 d Cognitive function Anxiety Ϫ0.359 .02 e Child 31 25.2 Ϯ 4.9 0.8 28.4 Ϯ 3.9 Depression Ϫ0.581a .00 Adolescent 19 27.3 Ϯ 3.0 0.1 27.6 Ϯ 4.1 HRQoL, health-related quality of life. Ϯ Ϯ Mean SD, mean standard deviation. aSignificant at p Ͻ .005. aAnxiety scores range from 20–60; higher scores represent more anxiety; bdepression scores range from 0–54; higher scores represent more depression; cone girl and one boy did not fill in the depression questionnaire; dcognitive function scores range from 0–32; higher scores represent better HRQOL; ep Ͻ .005 study group vs. norm data. chological function, and cognitive func- tion of children surviving septic shock. Table 3. Patient characteristics of children with and without self-reported cognitive problems HRQoL, anxiety, and depression scores of children who survived septic shock are on Children With Cognitive Children Without Cognitive average better or similar to the scores of Problems (n ϭ 20) Problems (n ϭ 25) children and adolescents of the same age Median (Range) Median (Range) p and gender in the general Dutch popula- tion. However, cognitive function, as re- Age of child at PICU admission, yrs 1.8 (0.0–7.0) 5.0 (0.1–17.0)a .00 Length of stay in PICU, days 6.5 (2.0–15.0) 4.0 (2.0–35.0) .19 ported by the children themselves, is Length of artificial ventilation, days 5.0 (0.0–14.0) 2.0 (0.0–25.0) .35 worse in septic shock survivors compared Risk of Mortality, PIM2, % 9.2 (1.1–26.2) 6.5 (1.0–26.1) .18 with the norm population. A younger age Age of child at follow-up study, yr 9.7 (8.1–13.0) 12.0 (8.0–20.4) .03 child at PICU admission is associated Highest creatinine in PICU, ␮mol/L 60.0 (18.0–227.0) 55.5 (26.0–296.0) .78 Highest lactate in PICU, mmol/L 3.1 (0.7–16.4) 3.0 (0.8–23.0) .60 with lower cognitive function. In addi- n (%) tion, a larger number of septic shock- surviving children, although not signifi- Female 10 (50.0) 9 (36.0) .35 Physical sequelae (yes) 9 (45) 6 (24) .16 cant, attend special education schools. Artificial ventilation (yes) 15 (75.0) 17 (68.0) .60 Studies in adult survivors of severe sepsis and septic shock showed HRQoL PICU, pediatric intensive care unit; PIM, Pediatric Index of Mortality. outcomes similar to those in other adults a p Ͻ .005 children with self-reported cognitive problems compared with children without self- surviving ICU admission, but lower than Ͻ reported cognitive problems. Children with cognitive problems scored 25% of the norm population. the general population (2). Studies focus- ing on HRQoL after surviving pediatric (t ϭ 1.172, df ϭ 11, p ϭ .266) did not ported worse cognitive function had septic shock are scarce (13, 14). In con- differ significantly from the general significantly more negative emotions, trast with our findings, HRQoL was de- Dutch population. more problems with their parents, and creased on the physical domains of these Risk Factors for Cognitive Problems. reported significantly more bullying, studies. The severity of illness and Children who reported cognitive prob- higher anxiety, and depression scores chronic complaints negatively affected lems were significantly younger at PICU compared with children who had better HRQoL (13, 14). One study focusing on admission compared with children with- cognitive function (Table 4). HRQoL in children surviving bacterial out cognitive problems (Table 3). A back- Follow-up Time. No statistically sig- meningitis showed a lower HRQoL com- ward regression analysis for cognitive nificant correlations were found between pared with the general population, espe- function showed a final model (R2 ϭ .13, follow-up time and HRQoL, anxiety, de- cially in children with worse cognitive F ϭ 6.5, p ϭ .015) with one significant pression, and cognitive function (data not function due to the meningitis (21). Be- risk factor: age of the child at PICU ad- shown). cause most effect sizes were small-to- mission (B ϭ 0.59, 95% confidence inter- moderate in this study, the clinical im- ϭ val, 0.1–1.1; p .015). DISCUSSION portance of these differences in HRQoL is Relationship Between Cognitive Func- relative. Similar to our findings, several tion and HRQoL, Anxiety, and Depres- This is one of the first follow-up stud- studies in pediatric oncology show better sion. Children and adolescents who re- ies to describe self-reported HRQoL, psy- or comparable HRQoL, anxiety, and de-

Pediatr Crit Care Med 2009 Vol. 10, No. 6 639 pression scores compared with the norm are disease-specific or present in PICU low-up studies in the PICU have had sim- population (35, 36). children in general, and examine risk fac- ilar response rates, this may have biased Few studies describe psychological ad- tors like hypoperfusion, age at time of our results. In addition, no statistically justment in PICU survivors after dis- illness, delirium, hypoxia, sepsis, and in- significant differences were found be- charge (16, 37). PICU admission is asso- flammation (20, 42). tween patient characteristics of partici- ciated with emotional, behavioral, and Because we did not include a control pants and nonparticipants. However, psychiatric symptoms. In contrast with group, it is not clear whether there is an there is a trend (p Ͻ .1) that the majority these findings, anxiety and depression association between cognitive problems of the participating children were scores in our study group were similar to and sepsis in our patients. We performed younger and admitted for a longer period or even better than normative data from another follow-up study on HRQoL in a compared with the nonparticipating chil- the Dutch population. Our results are small heterogeneous group of PICU sur- dren. This could have resulted in a selec- difficult to compare with these earlier vivors. No cognitive impairment was tion bias by excluding the older cases studies, considering the differences in found in 27 children (aged 8–15 yrs) ad- admitted for a shorter stay. Second, the follow-up time. mitted for a broad range of diseases (e.g., number of studied children is relatively In short, HRQoL, anxiety, and depres- asthma, trauma, meningococcal disease, small, and due to different ages, gender, sion in these septic shock survivors are cardiac failure). Because of the small and follow-up times, strong conclusions equal to or slightly better than in the number of children in this study, we are difficult. Therefore, a case control age-related Dutch norm population. Pa- could not analyze differences in cognitive study matched on age, gender, follow-up tients confronted with a life-threatening functioning in children admitted with time, and hospital or PICU admission disease are faced with the necessity to different diseases (12). Our studies and would be interesting. Third, seven chil- accommodate to the illness. Two mecha- the study of Elison et al suggest that dren were excluded because they did not nisms describing adaptation to stressful children with septic shock are at risk for complete the Dutch questionnaire ade- events give a possible explanation for this developing cognitive problems (40). quately due to language barrier. Our re- positive outcome: 1) the concept of post- In our study, cognitive function prob- sults therefore cannot be extrapolated to traumatic growth, defined as “the experi- lems were especially reported by children all ethnic and cultural groups living ence of significant positive change arising who were younger at the time of PICU in the Amsterdam region. Fourth, in the from the struggle with a major life cri- admission. Follow-up studies after pedi- present study, cognitive function was sis”; and 2) the concept of response shift, atric brain injury and in animal studies measured with a short self-report ques- defined as “the experience of hardship indicate that the younger brain may be tionnaire. Yet, diagnostic clarification changing the internal standard of pa- more vulnerable for the development of and grading of clinical severity of cogni- tients, resulting in changes in the mean- cognitive problems (43, 44). Possibly tive disorders should be evaluated by ing of self-evaluation and hence in a pos- other risk factors, such as meningitis and comprehensive neuropsychological test- sibly different experience of problems and pathophysiologic mechanisms mediating ing. These tests are designed to examine a values” (38, 39). These two mechanisms cognitive dysfunction, need to be ana- variety of cognitive abilities, including describe a process of adaptation to stress- lyzed in larger populations in future speed of information processing, atten- ful events, leading to a possible change studies (20). tion, memory, language, and executive for better-perceived HRQoL and de- Cognitive function problems, as eval- functions. creased scores for anxiety and depression uated by the children themselves, are as- In addition, younger children (Ͻ8 yrs questionnaires. We do not know how pa- sociated with more depressive moods, of age) were not evaluated. Future re- rental perceptions of the child surviving bullying, and depression in the current search should evaluate cognitive function influences children’s psychological func- study despite the fact that HRQoL and in all pediatric septic shock survivors tion. More detailed studies focusing on depression in the whole study group were with standardized neuropsychological these mechanisms are needed in children comparable with the norm population. tests. Besides, premorbid health status is surviving critical illnesses. Children with cognitive function prob- likely an important factor but is difficult Almost half (44%) of the children lems may adapt less well to stressful to assess (49). Finally, the definition of studied reported cognitive function prob- events and may be more vulnerable to children as being “at risk” for a cognitive lems and 14% of the children attended other problems than children without problem was based on the value of the special education schools. Survivors of cognitive function problems. Prospective 25th percentile of the domain in the norm septic shock seem to be at risk for adverse follow-up studies with adequate neuro- population. Because there is no gold neurodevelopmental outcomes also affect- psychological testing are necessary to standard comparison, the cutoff point ing academic performance. These find- evaluate cognitive function and risk fac- may seem arbitrary. This method, how- ings are in accordance with a recent pilot tors (before, during, and after PICU ad- ever, compared with contrasting means, study on neuropsychological function in mission) for cognitive function problems reveals clear differences between our children following PICU admission (40). (45, 46). Awareness of long-term sequelae study group and healthy controls. Results suggest impaired memory and at- may result in supportive programs after Despite these limitations, this is one of tention in these children, and a specific discharge, as is the case for neonatal and the few studies that provides insight into deficit in children with septic illness. trauma patients (47, 48). psychosocial behavior of pediatric septic Also, earlier reports on survivors of trau- A number of limitations to this study shock survivors. We found that the ma- matic brain injury, bacterial meningitis, should be taken into account. First, this jority of these survivors matches or even or meningococcal disease show neuro- is a retrospective study in one center with surpasses the general population in logic deficits (19, 21, 41). Future research no control group. The response rate of HRQoL, anxiety, and depression scores, should consider whether these deficits our study was 77%. Although other fol- and cognitive problems. These findings

640 Pediatr Crit Care Med 2009 Vol. 10, No. 6 are noteworthy and require further dis- 3. Jones C, Griffiths RD, Slater T, et al: Signif- long term after septic shock caused by Neis- cussion and research. icant cognitive dysfunction in non-delirious seria meningitides. Br J Clin Psychol 2008; patients identified during and persisting fol- 47:251–263 lowing critical illness. Intensive Care Med 18. Vermunt LC, Buysse CM, Joosten KF, et al: Suggestions for Future 2006; 32:923–926 Self-esteem in children and adolescents after Research 4. Needham DM, Dowdy DW, Mendez-Tellez septic shock caused by Neisseria meningiti- PA, et al: Studying outcomes of intensive dis: Scars do matter. J Adolesc Health 2008; Cohort and case-control studies of care unit survivors: Measuring exposures and 42:386–393 PICU survivors evaluating patient out- outcomes. Intensive Care Med 2005; 31: 19. Fellick JM, Sills JA, Marzouk O, et al: Neu- comes (physical, psychological, and cog- 1153–1160 rodevelopmental outcome in meningococcal nitive sequelae, and quality of life) and 5. World Health Organization: The First Ten disease: A case-control study. Arch Dis Child risk factors for sequelae (before, during, Years of the World Health Organization. Ge- 2001; 85:6–11 and after PICU admission) are essential. neva, Switzerland, World Health Organiza- 20. Herridge MS, Batt J, Hopkins RO: The patho- Awareness of long-term sequelae and tion, 1979 physiology of long-term neuromuscular and their risk factors may result in changes in 6. Rajmil L, Herdman M, Fernandez De San- cognitive outcomes following critical illness. treatment during the acute phase and in mamed MJ, et al: Generic health-related Crit Care Clin 2008; 24:179–199 quality of life instruments in children and 21. Koomen I, Raat H, Jennekens-Schinkel A, et supportive programs after discharge (50, adolescents: A qualitative analysis of content. al: Academic and behavioral limitations and 51). Long-term follow-up clinics of PICU J Adolesc Health 2004; 34:37–45 health-related quality of life in school-age survivors and rehabilitation programs 7. Berra S, Ravens-Sieberer U, Erhart M, et al: survivors of bacterial meningitis. Qual Life comparable with follow-up care in neona- Methods and representativeness of European Res 2005; 14:1563–1572 tal and trauma patients should be devel- surveys in children and adolescents: The 22. Nguyen DN, Spapen H, Su F, et al: Elevated oped to detect, support, and treat chil- KIDSCREEN study. BMC Public Health serum levels of S-100beta protein and neu- dren with cognitive, developmental, and 2007; 7:182 ron-specific enolase are associated with brain psychological problems. These programs 8. Ambuehl J, Karrer A, Meer A, et al: Quality of injury in patients with severe sepsis and sep- are expected to improve daily life and life of survivors of paediatric intensive care. tic shock. Crit Care Med 2006; 34:1967–1974 minimize the impact on children’s well- Swiss Med Wkly 2007; 137:312–316 23. Sharshar T, Carlier R, Bernard F, et al: Brain being and future development (47, 48). 9. Jones S, Rantell K, Stevens K, et al: Outcome lesions in septic shock: A magnetic reso- at 6 months after admission for pediatric nance imaging study. Intensive Care Med intensive care: A report of a national study of 2007; 33:798–806 CONCLUSION pediatric intensive care units in the United 24. Goldstein B, Giroir B, Randolph A: Interna- kingdom. Pediatrics 2006; 118:2101–2108 tional pediatric sepsis consensus conference: In this group of septic shock survivors, 10. Morrison AL, Gillis J, O’Connell AJ, et al: Definitions for sepsis and organ dysfunction HRQoL, anxiety, and depression were bet- Quality of life of survivors of pediatric inten- in pediatrics. Pediatr Crit Care Med 2005; ter or equal compared with the age- sive care. Pediatr Crit Care Med 2002; 3:1–5 6:2–8 related Dutch norm population. Appar- 11. Taylor A, Butt W, Ciardulli M: The functional 25. Levy MM, Fink MP, Marshall JC, et al: 2001 ently, children are able to overcome the outcome and quality of life of children after CCM/ESICM/ACCP/ATS/SIS International stressful experience of septic shock. Chil- admission to an intensive care unit. Inten- Sepsis Definitions Conference. Crit Care Med dren admitted at a younger age especially sive Care Med 2003; 29:795–800 2003; 31:1250–1256 reported cognitive function problems. 12. Knoester H, Bronner MB, Bos AP, et al: Qual- 26. The KIDSCREEN Group Europe. The KID- ity of life in children three and nine months SCREEN Questionnaires—Quality of life Prospective follow-up studies with ade- after discharge from a paediatric intensive questionnaires for children and adolescents. quate neuropsychological testing and in- care unit: a prospective cohort study. Health Lengerich, Germany, Lengerich:Pabst Sci- formation on risk factors, such as men- Qual Life Outcomes 2008; 6:21 ence Publishers, 2006 ingitis, are warranted to evaluate the 13. Buysse CM, Raat H, Hazelzet JA, et al: Long- 27. Spielberger CD, Edwards CD, Lushene LE, et association between septic shock, cogni- term health-related quality of life in survi- al: Manual for State-Trait Anxiety Inventory tive function, and risk factors for cogni- vors of meningococcal septic shock in child- for Children. Palo Alto, California, Consult- tive problems. hood and their parents. Qual Life Res 2007; ing Psychologist Press, 1973 16:1567–1576 28. Bakker FC, van Wieringen PCW, van der ACKNOWLEDGMENT 14. Buysse CM, Raat H, Hazelzet JA, et al: Sur- Ploeg HM, et al: Handleiding bij de Zelf- viving meningococcal septic shock: health beoordelings-vragenlijst voor kinderen (ZBV-K), We thank N. Peek, PhD, Department consequences and quality of life in children een Nederlandstalige bewerking van de of Medical Informatics, Academic Med- and their parents up to 2 years after pediatric State-trait anxiety inventory for children intensive care unit discharge. Crit Care Med (STAIC) van Spielberger e.a. (ZBV-K/STAIC). ical Center, for advice and statistical 2008; 36:596–602 Lisse, Netherlands, Swets en Zeitlinger, 1989 expertise. 15. Bronner MB, Knoester H, Bos AP, et al: Post- 29. Kovacs M: Children’s Depression Inventory. traumatic stress disorder (PTSD) in children New York, Multi Health Systems, 1992 REFERENCES after paediatric intensive care treatment 30. Timbremont B, Braet C: Children’s Depres- compared to children who survived a major sion Inventory, Handleiding, Nedreland- 1. Kutko MC, Calarco MP, Flaherty MB, et al: fire disaster. Child Adolesc Psychiatry Ment stalige versie. Lisse, Netherlands, Swets and Mortality rates in pediatric septic shock with Health 2008; 2:9 Zeitlinger Test Publishers, 2002 and without multiple organ system failure. 16. Shears D, Nadel S, Gledhill J, et al: Psychi- 31. Verrips EGH, Vogels TGC, Koopman HM, et Pediatr Crit Care Med 2003; 4:333–337 atric adjustment in the year after meningo- al: Measuring health-related quality of life in 2. Granja C, Dias C, Costa-Pereira A, et al: Qual- coccal disease in childhood. J Am Acad Child a child population. Eur J Public Health 1999; ity of life of survivors from severe sepsis and Adolesc Psychiatry 2007; 46:76–82 9:188–193 septic shock may be similar to that of others 17. Vermunt LC, Buysse CM, Joosten KF, et al: 32. Vogels AGC, Bruil J, Stuifbergen M, et al: who survive critical illness. Crit Care 2004; Behavioural, emotional, and post-traumatic Validity and reliability of a generic health- 8:R91–R98 stress problems in children and adolescents, related quality of life instrument for adoles-

Pediatr Crit Care Med 2009 Vol. 10, No. 6 641 cents, the TACQOL. Qual Life Res 1999; cations of findings from studies of response treatment. Ment Retard Dev Disabil Res Rev 8:630 shift. Health Qual Life Outcomes 2004; 2:14 2006; 12:168–173 33. Grootenhuis MA, Koopman HM, Verrips EG, 40. Elison S, Shears D, Nadel S, et al: Neuropsy- 46. Gordon SM, Jackson JC, Ely EW, et al: Clin- et al: Health-related quality of life problems chological function in children following ad- ical identification of cognitive impairment in of children aged 8–11 years with a chronic mission to paediatric intensive care: A pilot ICU survivors: Insights for intensivists. In- disease. Dev Neurorehabil 2007; 10:27–33 investigation. Intensive Care Med 2008; 34: tensive Care Med 2004; 30:1997–2008 34. Cohen J: Statistical Power Analysis for the 1289–1293 47. McCormick MC, Brooks-Gunn J, Buka SL, et Behavioral Sciences. New York, Academy 41. van Heugten CM, Hendriksen J, Rasquin S, al: Early Intervention in Low Birth Weight Press, 1988 et al: Long-term neuropsychological perfor- Premature Infants: Results at 18 Years of Age 35. Langeveld NE, Stam H, Grootenhuis MA, et mance in a cohort of children and adoles- for the Infant Health and Development Pro- al: Quality of life in young adult survivors of cents after severe paediatric traumatic brain gram. Pediatrics 2006; 117:771–780 childhood cancer. Support Care Cancer injury. Brain Inj 2006; 20:895–903 48. Savage RC, Pearson S, McDonald H, et al: 2002; 10:579–600 42. Hopkins RO, Ely EW, Jackson JC: The role of After hospital: working with schools and 36. Zebrack BJ, Zevon MA, Turk N, et al: Psycho- future longitudinal studies in ICU survivors: families to support the long term needs of logical distress in long-term survivors of Understanding determinants and pathophys- children with brain injuries. NeuroRehabili- solid tumors diagnosed in childhood: A re- iology of brain dysfunction. Curr Opin Crit tation 2001; 16:49–58 port from the childhood cancer survivor study. Pediatr Blood Cancer 2007; 49:47–51 Care 2007; 13:497–502 49. Demellweek C, Baldwin T, Appleton R, et al: 37. Small L, Melnyk BM: Early predictors of 43. Anderson V, Catroppa C, Morse S, et al: A prospective study and review of pre-morbid post-hospital adjustment problems in criti- Functional plasticity or vulnerability after characteristics in children with traumatic cally ill young children. Res Nurs Health early brain injury? Pediatrics 2005; 116: brain injury. Pediatr Rehabil 2002; 5:81–89 2006; 29:622–635 1374–1382 50. Dowdy DW, Needham DM, Mendez-Tellez 38. Barakat LP, Alderfer MA, Kazak AE: Posttrau- 44. Barichello T, Martins MR, Reinke A, et al: PA, et al: Studying outcomes of intensive matic growth in adolescent survivors of can- Cognitive impairment in sepsis survivors care unit survivors: The role of the cohort cer and their mothers and fathers. J Pediatr from cecal ligation and perforation. Crit Care study. Intensive Care Med 2005; 31:914–921 Psychol 2006; 31:413–419 Med 2005; 33:221–223 51. Knoester H, Grootenhuis MA, Bos AP: Out- 39. Rapkin BD, Schwartz CE: Toward a theoret- 45. Armstrong FD: Neurodevelopment and come of paediatric intensive care survivors. ical model of quality-of-life appraisal: Impli- chronic illness: Mechanisms of disease and Eur J Pediatr 2007; 166:1119–1128

642 Pediatr Crit Care Med 2009 Vol. 10, No. 6