Methods

Children discharged from PICU with TBI from 1999 to 2004 were identified from the PICU admission records. This period was chosen to allow sufficient time for outpatient review and growth evaluation. Children aged <1 year at injury were excluded due to inaccuracies in measuring this age group. Documentation and plotting of growth measurements were identified from case notes. The Glasgow Coma Scale (GCS) score on admission was recorded as an index of TBI severity.

Height and weight standard deviation scores (SDS) were calculated from British 1990 Growth Reference data.6 Height velocity SDS (HVSDS) was calculated in patients with measurements separated by 6–18 months.

Data were analysed using SPSS v 15.0. Patients were divided according to type of outpatient review and compared using Fisher’s exact test. The Wilcoxon signed rank test was used to compare height or weight SDS. Data are presented as mean (SD) unless otherwise stated.

Local research ethics approval was not required as this study was considered to be service evaluation.

Results

A total of 129 children aged >1 year were discharged from PICU following TBI during 1999–2004. Case notes for 123 were available for review. Demographic data and GCS score at presentation are shown in table 1.

Outpatient review following TBI

Following TBI, 85 of the 123 (69%) children were reviewed in outpatient clinics at the regional centre. This was represented by 212 clinic attendances, of which 192 were follow-up of neurological, orthopaedic or maxillofacial injuries (table 2). Twenty attendances (9%) were to review incidental findings or unrelated medical conditions. Twenty one children were seen in two or more specialty clinics.

Children undergoing outpatient review did not differ significantly from those not followed-up in terms of gender, age or GCS score on admission (table 1). Other recorded indices of TBI severity were similar between the two groups. Children not reviewed locally were more likely to have been transferred to another inpatient unit (p = 0.017). Neurology and/or neurosurgery clinics reviewed the largest number of children (53%) with the first review at 6.0 (SD 6.0) months with discharge at 12.0 (SD 13.2) months after TBI.

Outpatient documentation of growth

Only 71 (33%) of the 212 clinic attendances had documented height and weight measurements, of which six (8%) had been plotted on a growth chart. This represents measurements for 38 different children. Time to first measurement was 8.4 (SD 7.2) months after TBI. Nine attendances documented a reason for difficulty in measuring growth, including the child being in plaster or in a wheelchair. A marked variation in the practice of growth monitoring between clinics was observed. Neurosurgical, orthopaedic and oral-maxillofacial clinics did not routinely measure any children (table 1).

Serial growth measurements

Fifty three patients attended an outpatient clinic on at least two occasions. Serial measurements were available for only 22 patients, of which 19 were performed in neurology outpatients. Age and TBI severity of this subgroup were similar to the whole cohort (table 2).

A pre-injury measurement was available for only seven patients who had previously been reviewed in the outpatient clinic. The time interval from this baseline measurement to TBI was 31.2 (SD 26.4) months. For patients without a pre-TBI measurement, the interval to the first measurement, used as baseline, was 5.9 (SD 4.8) months. Mean duration of follow-up was 25.2 (SD 21.6) months after TBI.

We found a small reduction in mean HSDS in patients with serial growth measurements. Height SDS (HSDS) at baseline was −0.37 (SD 0.79) and had reduced significantly at the last documented measurement by 0.17 (SD 0.33) (p = 0.017). There were trends to increasing weight SDS (−0.20 (SD 1.00) to 0.15 (SD 1.32), p = 0.077) or body mass index SDS (0.14 (SD 0.99) to 0.52 (SD 1.29), p = 0.072). No formal assessment of mobility was made, but all children were independently mobile.

No child was investigated for poor linear growth. However, this was limited by the number of data points (median 3, range 2–11). HVSDS could be calculated in 16 cases, of which four were greater than 2 SD below the mean. Data were not available to determine if low HVSDS was sustained in these patients.

Discussion

Our evaluation of growth monitoring following TBI has shown this practice was poorly performed in this cohort. Growth documentation cannot be relied upon to detect all children with PTHP but may identify those requiring further investigation.

Our data demonstrate that following TBI, children are reviewed in various specialist clinics. Serial growth measurements were rarely available. This represents failure to perform growth monitoring but is also limited by the relatively short duration of neurological follow-up and half the children being discharged from specialist follow-up at the first outpatient review. However, a number of children may be followed up at local centres. This highlights the need for the development of protocols to ensure growth monitoring occurs, even in children discharged with apparent full recovery. Collaboration between tertiary centres, local hospitals and primary care is needed to ensure height is documented prior to discharge from hospital and on a regular basis thereafter. Currently, these data may be held in case notes at each centre. As such, centralisation of the data may be required to ensure early detection and prompt investigation of those with poor growth trends. Further research may also identify injury and patient characteristics which are associated with a higher likelihood of PTHP to allow targeted monitoring of high risk patients.

There are limitations to the interpretation of our data. Firstly, this is a highly selected subpopulation who required extended medical review, and relies on the accuracy of the documented measurements. Additionally, mid-parental height and pubertal status were not documented but would have allowed more accurate interpretation relative to height potential. Moreover, central precocious puberty and hypothalamic obesity are more common in children following TBI, and in these individuals GHD may not be detected by growth monitoring alone.

None of the patients met strict criteria to initiate investigation for GHD.7 However, application of these criteria was limited by either infrequent measurements or short duration of follow-up. HVSDS was markedly reduced in four cases, further highlighting the need for ongoing growth evaluation by primary or secondary care following discharge from tertiary centre review.

Conclusions

Despite the potential use of growth monitoring as a marker of pituitary function, growth has not been accurately documented in outpatient clinics following TBI to date. A co-ordinated approach by specialist and primary care is needed to ensure growth monitoring is performed until a more detailed consensus on the investigation of PTHP in childhood is reached.