SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Participant categorization
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Disclosure of financial interests
  10. References
  11. Supporting Information

Aim  The aim of this study was to determine the yield of magnetic resonance imaging (MRI) after an episode of childhood convulsive status epilepticus (CSE) and to identify the clinical predictors of an abnormal brain scan.

Method  Children were recruited following an episode of CSE from an established clinical network in north London. Eighty children (age range 1mo–16y; 39 males; 41 females) were enrolled and seen for clinical assessment and brain MRI within 13 weeks of suffering from an episode of CSE. Scans were reviewed by two neuroradiologists and classified as normal (normal/normal-variant) or abnormal (minor/major abnormality). Factors predictive of an abnormal scan were investigated using logistic regression.

Results  Eighty children were recruited at a mean of 31.8 days (5–90d) after suffering from CSE. Structural abnormalities were found in 31%. Abnormal neurological examination at assessment (odds ratio [OR] 190.46), CSE that was not a prolonged febrile seizure (OR 77.12), and a continuous rather than an intermittent seizure (OR 29.98) were all predictive of an abnormal scan. No children with previous neuroimaging had new findings that altered their clinical management.

Interpretation  Brain MRI should be considered for all children with a history of CSE who have not previously undergone MRI, especially those with non-prolonged febrile seizure CSE, those with persisting neurological abnormalities 2 to 13 weeks after CSE, and those with continuous CSE.


Abbreviations
CSE

Convulsive status epilepticus

GOSH

Great Ormond Street Hospital

HIMAL

Hippocampal malrotation

PFS

Prolonged febrile seizure

STEPIN

Status Epilepticus Imaging and Neurocognitive Study

What this paper adds

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Participant categorization
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Disclosure of financial interests
  10. References
  11. Supporting Information
  •  A significant proportion of children will have structural abnormalities on MRI following CSE.
  •  Clinical factors such as persistent neurological abnormality, continuous CSE, and non-prolonged febrile seizure-related CSE can potentially be used to select children at increased risk of having an underlying structural abnormality for brain MRI.
  •  CSE alone is probably an insufficient reason for repeating MRI in children with known structural brain lesions or who have had previous normal MRI.

Convulsive status epilepticus (CSE) in children is common1–3 and potentially life-threatening,1,4 with high associated morbidity. Guidelines on acute management of CSE5 focus on drug therapy6 and seizure termination, but guidance on optimal follow-up investigations remains unclear.7 In particular, one investigation frequently considered is the need for and type of neuroimaging. Since the outcome is largely dependent on the aetiology,1,8 determining the underlying diagnosis is important for both treatment and prognosis, and neuroimaging may thus be useful.

From historical data, a minimum yield of detectable lesions of 7.8% has been estimated amongst all children with CSE who underwent neuroimaging.7 In practice, this may be conservative, particularly if imaging is restricted to children at particular risk of structural brain lesions. It would be beneficial to determine yield in the general population of children with CSE and identify risk factors for structural lesions so that unnecessary imaging is avoided and those children likely to benefit are treated appropriately.

The Status Epilepticus Imaging and Neurocognitive Study (STEPIN) is a prospective study on the effects of childhood CSE within the first year of suffering from an episode of CSE. In this paper, we report on the clinical and magnetic resonance imaging (MRI) findings within 13 weeks of CSE and on factors predictive of abnormal scans. These data are needed to inform policy and guideline development for the management of childhood CSE.

Method

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Participant categorization
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Disclosure of financial interests
  10. References
  11. Supporting Information

In STEPIN, children presenting to north London hospitals with CSE (defined as a convulsive seizure or series of seizures lasting more than 30 minutes without recovery of consciousness) between 1 March 2007 and 1 March 2010 were identified using an established clinical/research network.1 This network consists of 18 hospitals with 24-hour paediatric accident and emergency services, five paediatric intensive care units, and the regional centralized paediatric intensive care retrieval service. Children with CSE were notified to a centralized research team by their admitting local paediatrician/care retrieval service. Each family was contacted and invited by the research team to participate. Eighty children (39 males; 41 females; age range 1mo–16y) were enrolled and invited to have a brain MRI at Great Ormond Street Hospital (GOSH) within 1 to 13 weeks after CSE.

The MRIs were performed where possible with the children either awake and lying still or in natural sleep. If this was judged impossible owing to the developmental status of the child then, with parental consent, sedative medication was offered to increase the chance of the child staying still during scanning.

The MRIs were carried out using the GOSH protocol for the evaluation of children with epilepsy and optimized for visualization of mesial temporal structures. It included a T1-weighted three-dimensional fast low-angle shot MRI sequence (repetition time 11ms; echo time 4.94ms; acquisition matrix 244 × 256; in-plane resolution 1.0 × 1.0mm; slice thickness 1mm), T2-weighted axial and coronal images (repetition time 6170ms; echo time 14ms; acquisition matrix 216 × 320; slice thickness 4mm), diffusion-weighted images (repetition time 2700ms; echo time 96ms; acquisition matrix 128 × 128; slice thickness 5mm), and T2 relaxometry (repetition time 2400ms; echo time six echoes ranging from 26–352ms; acquisition matrix 150 × 256; in-plane resolution 2.5 × 2.5mm; slice thickness 5mm). All images were acquired on a 1.5T Siemens Avanto MRI system (Siemens AG, Muenchen, Germany). Sequences were also performed for diffusion tensor imaging. However, as this is not currently in routine clinical use in children, this was not included in the evaluation performed for the current paper.

Participant categorization

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Participant categorization
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Disclosure of financial interests
  10. References
  11. Supporting Information

When individuals attended GOSH for their MRI, a clinical history was taken from each family and a neurological examination was performed (MY). Information obtained included the child’s medical history, the results of investigations performed during his or her acute hospital admission, seizure semiology, and any concerns about development before CSE. Seizure duration was estimated from the initial referral report and corroborated by parental recollection. All information (including results of previous neuroimaging before enrolment) was collated and used by two paediatric neurologists (RC and RS) blinded to results of the current study MRIs to categorize children into one of seven aetiological groups: prolonged febrile seizure (PFS); acute symptomatic; remote symptomatic; acute on remote symptomatic; idiopathic epilepsy-related; cryptogenic epilepsy-related; and other CSE as defined previously.1 Clinical data were assessed independently and differences in opinion were resolved by consensus.

Scan categorization

Each MRI was evaluated by two paediatric neuroradiologists (MC and WC) for abnormalities and assigned to one of four groups: normal (no abnormal features); normal-variant (unusual feature, thought to be variation of normal with no functional significance); minor abnormality (abnormal feature thought to be either unrelated to this CSE episode/no functional significance); or major abnormality (abnormal feature likely to have significant impact on the child/represent a cause for this CSE episode). MRIs were also assessed for presence of hippocampal malrotation (HIMAL),9 which is associated with seizures.10,11 Evaluations were performed without knowledge of the children’s clinical histories.

Statistical analysis

All data were analysed using PASW Statistics 18.0.2 (Chicago, IL, USA). Logistic regression was used to investigate candidate variables predictive of an abnormal MRI. As even minor abnormalities may have clinical relevance to the individual child and in order to maximize our model’s predictive power, major and minor abnormalities were combined for analysis, as were normal and normal-variant categories. Based on previously reported clinical and demographic factors which may increase the risk for structural brain lesions,12,13 the following were investigated: age at CSE; seizure duration greater than 60 minutes; seizure focality; continuous vs intermittent CSE; seizure aetiology (PFS vs non-PFS); previous history of seizures or CSE; abnormal neurological examination at assessment; history of developmental delay pre-CSE; and results of previous neuroimaging. These were entered into a multivariable logistic regression analysis. Bootstrapping techniques were used to assess internal validity.14.

STEPIN was approved by the GOSH Local Research Ethics Committee. Written informed consent to participate in STEPIN and for each procedure involved was obtained from each enrolled individual or their family at the initial assessment.

Results

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Participant categorization
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Disclosure of financial interests
  10. References
  11. Supporting Information

Demographics

Of the 225 children who were notified to STEPIN, 80 children were enrolled and underwent MRI. Children were seen for assessment at a mean of 31.78 days after CSE (range 5–9d). Of those who were not enrolled, 35 children were uncontactable owing to missing or incorrect contact details; 38 children were unsuitable for MRI under sedation owing to instability of their clinical condition/comorbidities; 48 children declined to participate; 15 children who lived distant to the study area were not willing to visit our centre; and seven children died during their acute hospital admission. A further two children agreed to participate but did not attend their appointments. The 145 children who were not enrolled were not seen for assessment and so consent to participate in the study was not taken. Therefore, only minimal demographic details were available and a comparison of clinical features of these children and those enrolled was not possible. Children who did not participate did not differ significantly in age from those who did (mean age 3y 2mo vs 3y 10mo; t-test; p=0.147) or sex (male:female ratio 39:41 vs 85:60; χ2; p=0.154).

The ages of children included in this study ranged from 1 month to 16 years (mean age 3y 3mo; median 1y 11mo). Overall, one-third of children had focal seizure onset, the majority presented with tonic–clonic seizures, and over half had continuous rather than intermittent CSE. Sixty per cent of children had previous seizures of any nature (including previous simple febrile seizures), although in only 26% had this been CSE (Table I).

Table I.   Seizure characteristics of convulsive status epilepticus (CSE)
CharacteristicsOverall cohort; n=80 (%)
Aetiology of CSE
 Prolonged febrile seizure33 (41.3)
 Acute symptomatic3 (3.8)
 Remote symptomatic7 (8.8)
 Acute on remote symptomatic14 (17.5)
 Idiopathic epilepsy-related8 (10.0)
 Cryptogenic epilepsy-related11 (13.8)
 Other4 (5.0)
Focal onset23 (29.1)
Continuous seizure46 (57)
Intermittent seizure54 (43)
Semiology
 Tonic15 (17.5)
 Clonic3 (3.8)
 Tonic–clonic62 (77.5)
Mean seizure duration, min (range)72.40 (30–265)
Seizure duration >60min36 (45)
History of developmental delay24 (30)
Abnormal neurological examination13 (16.3)
Previous seizures48 (60)
Previous episode CSE21 (26)

MRI results

Overall, 25 of the 80 (31.2%) individual scans showed abnormal features. These were evenly split between major and minor abnormalities, as defined earlier. All major abnormalities were found in children with acute, remote, or acute on remote symptomatic CSE, and only one child with PFS showed a minor abnormality. A full breakdown of findings by aetiological group is in Table II (specific abnormalities found are given in Table SI, supporting material online). One child with PFS and no children with other forms of CSE met criteria for unilateral HIMAL whilst a further three children with PFS and one child with an unclassified episode of CSE met the partial criteria. Including these children within the diagnosis, the proportion of children with HIMAL was significantly higher in the group with PFS than in the other groups (9.1% vs 2.1%; p=0.029; χ2).

Table II.   Magnetic resonance imaging findings by aetiology of convulsive status epilepticus (CSE)
 Normal (%)Normal variant (%)Minor abnormality (%)Major abnormality (%)
Prolonged febrile seizure (n=33)30 (90.9)2 (6.1)1 (3.0)0
Acute symptomatic CSE (n=3)1 (33.3)01 (33.3)1 (33.3)
Remote symptomatic CSE (n=7)002 (28.6)5 (71.4)
Acute on remote symptomatic CSE (n=14)4 (28.6)03 (21.4)7 (50.0)
Idiopathic epilepsy-related CSE (n=8)7 (87.5)01 (12.5)0
Cryptogenic epilepsy-related CSE (n=11)7 (63.6)04 (36.4)0
Other CSE (n=4)4 (100)000
Total (n=80)53 (66.3)2 (2.5)12 (15.0)13 (16.2)

Logistic regression

Since not all children had previously undergone computed tomography (CT) or MRI, previous neuroimaging was not included in the regression model. Regression analysis revealed abnormal neurological examination, continuous CSE, and non-PFS aetiology as factors predictive of an abnormal MRI. Ten repeated bootstrap analyses using 1000 samples each consistently identified these three factors as significant, whereas others were not. Odds ratios, B-values, and their 95% confidence intervals (CIs) are reported in Table III.

Table III.   Clinical factors predictive of abnormal magnetic resonance imaging (MRI) after convulsive status epilepticus (CSE)
Predictive factorAbnormal MRI (%)p-valueB95% confidence interval (B)Odds ratio
  1. PFS, prolonged febrile seizure.

Abnormal neurological examination at the time of MRI12 (92.3)0.0015.2493.959 to 424.134190.460
Non-PFS/unclassified vs PFS24 (51.1)0.001–0.0024.3452.194 to 316.92377.108
Continuous vs intermittent CSE19 (40.4)0.001–0.0043.3991.240 to 390.45029.947
Developmental delay before CSE15 (62.5)0.442–0.4950.682−40.391 to 47.4391.978
Previous seizures18 (37.5)0.716–0.786−0.053−50.912 to 97.6620.949
Previous CSE9 (42.9)0.114–0.140−1.545−145.027 to 16.2450.248
Focal vs generalized onset9 (39.1)0.854–0.904−0.020−52.063 to 45.5510.980
Seizure duration >60min vs duration <60min8 (22.2)0.032–0.053−1.584−162.625 to 1.1340.205
Age0.299–0.357−0.135−40.097 to 4.9860.313

Previous neuroimaging

Of the 80 children, 30 (37.5%) underwent head CT during their acute admission with CSE. The findings were abnormal in 7 (23.3%). Four of these children also had an abnormal follow-up MRI. Lesions which appeared to resolve included cerebral oedema, white matter asymmetry, and wide subdural spaces. Of the 23 children with a normal acute CT 4 (17.4%) had an abnormal MRI – three minor and one major abnormalities. Of the 49 (34.7%) children who did not undergo CT, 17 had an abnormal MRI (Fig. S1, supporting material online). An abnormal CT was associated with an abnormal MRI (odds ratio 8.89, 95% CI 1.29–61.06; p=0.026).

Of the children in our cohort, 27 had previously undergone brain MRI before entry into STEPIN. This had been performed for a variety of indications, including investigation of developmental delay, previous unprovoked seizures, and acute meningitis. Repeat imaging after CSE allowed detection of five minor abnormalities previously not reported (Fig. S2, supporting material online), although they were present on the original MRIs performed before CSE, that is, these were old lesions (consistent with remote symptomatic aetiology) that were not new emergent findings and did not necessitate changes in treatment. No new major abnormalities were found, although one individual previously diagnosed as having cortical dysplasia was diagnosed as having tuberous sclerosis.

Discussion

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Participant categorization
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Disclosure of financial interests
  10. References
  11. Supporting Information

It is known from previous studies that have used a mixture of CT and MRI that a high proportion of children with CSE have abnormal neuroimaging. However, ours is the first study to follow a cohort with complete MRI coverage and to systematically investigate clinical factors associated with an abnormal MRI. The main findings are that (1) in over 30% of children brain MRI abnormalities will be detectable a median of 1 month following CSE and (2) clinical features can predict those children who are most likely to have structural abnormalities on MRI if it is performed as part of their follow-up investigations.

Our study found abnormalities on brain MRI in 31.3% of children with CSE. This is similar to the findings of a study of 144 children with new-onset CSE,13 which reported a 30% yield of neuroimaging abnormalities based on combined CT/MRI. Previous studies have reported that imaging children after a first unprovoked seizure has a yield of brain abnormalities of 13 to 32%,15,16 not accounting for seizure duration. As a high proportion of CSE is associated with an acute provoking factor,12 with the resultant possibility of associated structural brain injury, it is consistent that the yield of neuroimaging in CSE is at the upper end of this range.

A recent meta-analysis to estimate the prevalence of incidental MRI findings in healthy people found overall rates of intracranial abnormalities of 2.7% (95% CI 1.57–4.08).17 Thus, the yield of abnormal features of 31% in our cohort of children with CSE is substantially higher than that expected in a population without CSE and is unlikely to represent an incidental finding.

The specific predictive factors for an MRI-detected structural brain abnormality identified in the current study were having (1) an abnormal neurological examination after CSE, (2) a non-PFS CSE, or (3) continuous CSE. Thus, in deciding whether a child should undergo MRI following CSE, we would suggest that these findings be considered. We found no evidence that seizure focality, previous seizures or CSE, or duration of CSE were associated with abnormalities on MRI within 6 weeks of suffering from CSE and therefore suggest that these should be considered less important in deciding whether MRI is appropriate. Since the incidence of CSE is greatest in infants and younger children,12 the majority of children with CSE are unlikely to tolerate MRI without some form of sedation or anaesthesia. These associated procedures carry a risk, 18 albeit a small one (to the child), although MRI itself is believed to carry minimal health risks. Thus, the risks and difficulties associated with brain MRI in children following CSE need to be carefully weighed against the probable yield. We found only one minor abnormality and no major abnormalities among 33 children with PFS at 13 weeks after CSE, giving an estimated yield of major abnormalities of 0 to 10.3% (95% CI). Thus, although children with PFS make up a substantial proportion of childhood CSE and there are concerns over their longer-term risk of hippocampal damage,19 our data suggest that the majority will not have any structural abnormalities on MRI during the immediate follow-up period, but we do acknowledge that our sample size was modest. Although previous reports have suggested an association between PFS and HIMAL,20 and even though we found a greater number of children meeting HIMAL criteria in our PFS group, the clinical implications remain uncertain.21

Our study did not consider emergency management of CSE, so we are unable to comment on the use of head CT in the emergency department. None of the children in our study required emergency neurosurgery, although there may have been children in whom this was an important possibility to exclude. CT performed in the acute situation appears to detect a number of acute abnormalities that do not persist, and, although a normal CT scan reduces the likelihood of an MRI abnormality, the predictive power of either a normal or an abnormal CT scan in isolation is similar to the clinical factors described earlier. Thus, a normal CT scan does not offer complete reassurance that there no pathology would be found on MRI.

Although CT is more widely available, to detect acute causes of seizures (e.g. meningoencephalitis, venous sinus thrombosis), MRI is now the preferred imaging modality for epilepsy7 owing to its increased sensitivity in picking up remote symptomatic causes such as cortical malformations. 22 In addition, MRI does not utilize ionizing radiation, to which children are more sensitive than adults.23 Although we found abnormalities in five children in whom previous MRI elsewhere was reportedly normal, their original pre-CSE MRIs, on re-evaluation by the research team, were not normal and the abnormalities seen on repeat scans were already present. This would suggest that, in a child with a known abnormality on MRI or with otherwise identified remote symptomatic CSE, an episode of CSE by itself would not be sufficient reason to repeat the MRI.

A limitation of STEPIN is that the cohort examined may not be representative of the complete population of children with CSE. Our cohort contained a lower proportion of children with acute symptomatic (3.8% vs 17.1%) or remote symptomatic (8.8% vs 16.5%) CSE than our epidemiological study of CSE1 and other epidemiological studies of CSE.4,24 One reason for this was difficulties in recruiting individuals who remained unstable during the time period after their episode of CSE and thus were not able to enter the study. Several children with severe neurodisability agreed to participate but were unable to safely tolerate the sedation necessary for MRI. The individuals included in our study therefore do not include those with the more severe medical conditions, and thus there is a minor selection bias. Nonetheless, as those excluded were among the children who were most likely to have structural abnormalities on neuroimaging, this makes our estimate of the yield conservative and inclusion of these children would not be expected to alter the overall conclusion. Seizure semiology was similar to that found in our previous north London study with similar proportions of focal onset (28.8% vs 34.7%), continuous CSE (57.5% vs 47.7%), and tonic (18.3% vs 13.1%), clonic (3.8% vs 1.1%), or tonic–clonic seizures (77.5% vs 85.8%). There was a smaller proportion of children with CSE lasting over 60 minutes (45% vs 60%), which may reflect changes in medical management over the intervening time period.

Our study is the largest reported longitudinal study of children with CSE of all causes. While more precise estimates of odds ratios would have been possible with additional recruitment, practical limits of distance and travel time meant that it was not possible to expand the study area to capture a larger cohort within the study period.

It is difficult to comment on how the MRI findings affected the clinical management of these children, as responsibility for this remained with the referring hospitals. Reports from each scan were provided to the responsible clinician looking after each child, and our classification system meant that any major abnormalities found would have implications for further management and prognosis if found in a previously imaging-naïve child. It is also important, however, to remember that a normal MRI may provide reassurance to families. In this study MRI was timed to take place a minimum of 1 week after the acute episode of CSE. While previous work has shown that acute changes such as hippocampal oedema are common following CSE,25,26 the clinical relevance of these is uncertain. Owing to resource constraints, it is rarely feasible in a clinical setting in the UK to perform acute MRI. The timing of MRI was chosen to reflect this and would not have detected any self-limiting acute changes.

In summary, we suggest that MRI should be considered in the follow-up of children after CSE as a sizable proportion of children will have structural brain abnormalities. The findings have the potential to provide important diagnostic and prognostic information as well as directing therapy. The children most likely to have an abnormal scan are those with a persistent abnormal neurological examination and those who have had continuous CSE, while children who can be identified as having had PFS are less likely to benefit from routine neuroimaging. This study on MRI within 13 weeks of CSE is unable to determine whether CSE may lead to structural changes in the brain in the longer term, such as mesial temporal sclerosis. Instead, further longitudinal study is needed.

Acknowledgements

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Participant categorization
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Disclosure of financial interests
  10. References
  11. Supporting Information

We are thankful to the medical and nursing staff who were instrumental in helping to recruit individuals. We are grateful to the children and their families who participated, without whom this study would not have been possible. RS and RC were responsible for the concept and design of this study. MM and MY acquired the data, which was analysed by MY, MM, RC, RS, RM, and KC. Statistical analysis was performed by MY, RC, and RS, with additional advice from Dr Martin King. MY, RC, and BN drafted all or parts of the article, which was reviewed by MY, MM, RM, CC, KC, BN, RC, and RC before submission.

Disclosure of financial interests

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Participant categorization
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Disclosure of financial interests
  10. References
  11. Supporting Information

This study was funded by the Wellcome Trust (Grant number: 060214/HC/RL/MW/kj). This work was undertaken at GOSH/University College London Institute of Child Health, which received a proportion of funding from the Department of Health’s National Institute for Health Research Biomedical Research Centres funding scheme. The Centre for Paediatric Epidemiology and Biostatistics also benefits from funding support from the Medical Research Council in its capacity as the Medical Research Council Centre of Epidemiology for Child Health. Dr Yoong, Dr Madari, Dr Martinos, Dr Clark, and Dr Chong report no disclosures. Dr Chin held a National Institute for Health Research Academic Clinical Lectureship and received travel grants from GlaxoSmithKline, Janssen-Cilag, Esai, UCB Pharma. Professor Neville has received support for attending meetings and to run a meeting from Sanofi Aventis. He has attended a consultation meeting for UCB. Dr Scott is supported by GOSH Children’s Charity and has received travel grants from GlaxoSmithKline, Janssen-Cilag, UCB Pharma, and SPL Ltd.

References

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Participant categorization
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Disclosure of financial interests
  10. References
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Participant categorization
  6. Results
  7. Discussion
  8. Acknowledgements
  9. Disclosure of financial interests
  10. References
  11. Supporting Information
FilenameFormatSizeDescription
DMCN_4215_sm_FigS1.png35KSupporting info item
DMCN_4215_sm_FigS2.png37KSupporting info item
DMCN_4215_sm_TableS1.pdf60KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.