Computed tomography stroke findings and population demographics at Pelonomi Hospital, Bloemfontein

Note: A selection of conference abstracts: RSSA/SASPI Paediatric Imaging Congress, 03–06 November 2016, Spier Estate, Stellenbosch, South Africa. Faculty collaborators: Professor Kassa Darge (Body Imaging, University of Pennsylvania, Philadelphia, USA), Professor Edward Lee (Thoracic Imaging, Harvard University, USA), Professor Beverley Newman (Cardiac Imaging, Stanford University, California, USA), Professor Kimberly Applegate (Image Gently and Body Imaging, Emory University, Atlanta, USA) and Professor Savvas Andronikou (Thoracic Imaging, University of Bristol, UK) supported by South African Paediatric Radiologists, co-ordinated by Dr Jaishree Naidoo, President of the African Society of Paediatric Imaging and Head of Division of Paediatric Radiology, Charlotte Maxeke Johannesburg Academic Hospital. This conference abstract is partially based on the following publication: Daffue K, Joubert G, et al. Computed tomography stroke findings and population demographics at Pelonomi Hospital, Bloemfontein. S Afr J Rad. 2016;20(1), a993. https://doi.org/10.4102/sajr.v20i1.993 Background: Stroke remains the highest cause of death in patients more than 50 years old in South Africa and the fourth highest cause of death overall. There is a paucity of information regarding this disease in Free State province.


Introduction
The effect of stroke worldwide and in South Africa is well known.It remains, however, the highest cause of death in patients more than 50 years of age in South Africa, and the fourth highest cause of death overall. 1 The impact on surviving patients, their families and support systems is considerable.To alter the course of this burden, it is first necessary to know the disease profile in the local population, which will aid in identifying reversible factors in those at risk.Similar sociodemographic and risk profile studies have been conducted in South Africa 2,3,4 but research of this nature is lacking in the Free State Province.In addition, literature evaluating the various time parameters influencing stroke imaging and patient treatment is also deficient in South Africa.

Imaging and interpretation
All patients were scanned by a GE Discovery HD 750 64-slice multidetector computed tomography (MDCT) machine.
The imaging findings were interpreted by the radiology registrar responsible for the CT list on the day of imaging and were reviewed by the main researcher before data collection.Findings were interpretated according to a standardised departmental imaging and reporting protocol (Appendix 1).The imaging protocol was adapted from the recommended algorithm of Wintermark et al., with the exclusion of magnetic resonance imaging (MRI) as it was not routinely available during the study period. 5In summary, this entailed rapid assessment for ischaemic versus haemorrhagic stroke and the use of computed tomography angiography (CTA) and perfusion studies where indicated.The Alberta stroke programme early CT (ASPECT) score was applied to all acute ischaemic stroke patients considered for thrombolysis. 6nterpretation style was also modelled on previous works and recommendations. 7

Data analysis
Results were summarised by frequencies and percentages (categorical variables) and means, standard deviations or percentiles (numerical variables).

Ethical consideration
Ethics approval was obtained from the Ethics Committee of the Faculty of Health Sciences of the University of the Free State prior to the study initiation date.As imaging and interpretation of results were performed according to existing departmental guidelines, informed consent was not needed prior to imaging.All data were noted confidentially by using the patients' hospital number as identification.Permission was also obtained from the chief medical officer of Pelonomi Regional Hospital.

Risk factors and demographics
A total of 174 patients were included in the study.Their mean age was 59 years (SD 14.6), ranging from 17 years -96 years.The study population was predominantly female (n = 93, 53.5%).The most prevalent risk factor was hypertension (83.7%), followed by smoking (20.5%) and diabetes (15.0%) (Table 1).

Stroke type and territories
The population group consisted of 67.8% ischaemic (n = 118) and 32.2% (n = 56) haemorrhagic stroke patients.The majority of them (77.3%) had middle cerebral artery (MCA) territory infarcts (n = 92).The vascular territories of ischaemic infarcts are demonstrated in Figure 1.Four patients had no signs of ischaemia or haemorrhage (3.4%).Because these patients had focal neurological deficits at the time of imaging, they were included in the ischaemic group.Of these four, one had progressive neurological deficit in hospital.Clinical notes could not be retrieved for the other three patients.

Time from symptom onset and scan time
The TSO was unknown for 72 (41.4 %) of the total number of patients.Of the remaining 102, the majority (82.4%) presented after 8 hours.
In the group with a known TSO and ischaemic strokes (n = 67), only five patients (7.5%) presented within the four-and-ahalf-hour therapeutic window for thrombolysis (Figure 2).Of these five, only one received intravenous recombinant tissue plasminogen activator (rtPA).Clinical records on the outcome of this patient were not available.One patient had an ASPECT score of 4. One fell out of the therapeutic window owing to inter-hospital delays, and two did not receive rtPA for reasons not stated in the clinical notes.
The median ORT was 61 min (range 18-1361).Coincidentally, the median was the same for the group with a known TSO < 8 hours.The median ART was 32 min (range 4-893).For the group presenting within 8 hours (n = 18), the median was 36.5 min (Table 2).Nearly two-thirds (63.1%) of the CT requests were received after normal working hours.

Discussion
The patient mean age of 59 years was similar to that found by identified hypertension as the leading risk factor (70% of patients), followed by current smoking (30%) and alcohol consumption (26%). 3The prevalence of diabetes in their study was similar to that in our population group (14%).
Assessing the management of such medical conditions and advocating lifestyle modification by reinforcing public awareness is pivotal in addressing stroke management.The prevalence of HIV infection could not be assessed, as this was inconsistently recorded; this was unfortunate as it would have been an interesting contribution to the study.Tipping et al. found in their 2007 study that of 1087 stroke patients, 6.1% were HIV infected.The majority of the 67 patients were younger than 46 years, and had ischaemic strokes (96%).Infectious meningitis or vasculitides (28%) and HIV vasculopathy (20%) were more common underlying aetiologies.The more traditionally accepted risk factors (hypertension, diabetes, etc.) were not significant risk factors in this young stroke group. 10HIV infection and antiretroviral treatment contribute to accelerated atherosclerosis and aneurysmal changes, thus increasing the risk of stroke.As found by Benjamin et al., it would be difficult to identify HIV as a stand-alone risk factor in our population group, considering the high prevalence of hypertension. 11e stroke distribution (ischaemic 67.8%, haemorrhagic 32.2%) is in line with that of other population groups in Africa, where the incidence of haemorrhagic stroke ranged between 29% and 57%. 3,9,12Owing to unretrievable clinical records, transient ischaemic attacks could not be ascertained in terms of a cause in three of the four patients previously mentioned with normal CT findings.As stated in our imaging    Source: American Heart Association.Available from: http://www.strokeassociation.org/idc/groups/heart-public/@wcm/@hcm/@gwtg/documents/downloadable/ucm_308277.pdfprotocol, these patients could have benefited from additional CT perfusion studies but, owing to technical difficulties at the time (non-functional injection pump), this was not possible.A routine MRI service was also not yet available.
Given that 52.4% of the patients with acute ischaemic strokes had ASPECT scores ≥ 7, but only 7.5% presented within fourand-a-half hours from symptom onset, expediting the referral and transport process would significantly increase the number of patients qualifying for rtPA.Understandably, there are contributing factors beyond our control, such as the distance from hospital at the time of symptom onset.It should also be noted that our institution has an ever-enlarging catchment area owing to loss of resources in peripheral hospitals.Inhospital delays are not unique to our situation, as other developing countries experience similar problems, which was evident from an article published by Gurav et al. in May 2015. 13ny factors can, however, be addressed as outlined in the Target stroke campaign manual released by the American Heart Association (AHA). 8These include early notification of emergency departments by emergency medical services (EMS), education of the EMS personnel on appropriate triage of patients eligible for intravenous treatment, expedited inhospital assessment and improved guidelines and protocols to minimise uncertainty by the treating physicians.Such guidelines, specifically designed for the South African setting, have already been established. 14The treating physician could also accompany appropriately selected patients to the scanner; this would shorten the door-to-needle time as drug administration could then commence as soon as a relevant ischaemic stroke pattern is identified.
Another recommendation is to make use of a 'single-call activation system'.This was implemented with good effect by Nolte et al. 15 at their institution in the form of an all-points alarm.The entire stroke team was notified simultaneously by means of an SMS-based system.This method reduced their median in-hospital delay time from 54 min to 35 min and increased the number of patients treated with rtPA by 24%.
As time intervals were deduced from the RIS, the time of arrival in hospital and 'door to physician time' were unknown.We therefore used the ORT, as previously defined under methodology.This is an obvious underestimation but nonetheless is still longer than recommended by international guidelines. 8Our median ORT was 61 min, with the recommended 'door to report time' being 45 min.
The median reporting time of 32 min should also be addressed, as the AHA recommendations permit 20 min (door-to-scan report time: 45 min; door-to-scan time: 25 min).
Comparison of our time parameters with AHA guidelines is depicted in Table 2.This period could be improved by clear interpretation protocols and increased awareness on utilising tools such as the ASPECT score.The effect of delayed reporting time on treatment is crucial.
When applicable, the referring physician was telephonically informed of the CT findings immediately.A formal report was only released after a protracted period, which could explain the wide interquartile ranges.Measuring these variables was not practically possible and could not be assumed.
The fact that almost two-thirds of the scan requests occurred after hours is not surprising, as a working day consists of 8 hours.It is also a reflection of functionality of the after-hours services.
Following the outcome of the present study, it is our intention to reconvene with our Internal Medicine Department about implementing the above-mentioned plans and determining further progress.

Limiting factors
The small study population prohibits further delineation of results.As all the steps for data collection are already in place, a larger follow-up study would address this issue.Because we only considered patients who underwent CT, the total number of strokes during the period could be underestimated.However, patients who did not receive imaging probably did not fall within a period in which medical intervention would alter the course of the disease process.Unavailability of the exact TSO onset for 72 of the 174 patients is an obvious limitation that could not be overcome.This once again reiterates the importance of public awareness in stroke management.The onset and management of hypertension and diabetes was not recorded, as this was not part of the initial study outcomes.

Conclusion
Our stroke population did not differ significantly from those documented in the rest of South Africa and Africa as a whole.Pre-and in-hospital delays contributed significantly to patient numbers qualifying for thrombolytic treatment.
TSO could not be determined for 72 (41.1%) of the total of 174 patients.

FIGURE 2 :FIGURE 3 :
FIGURE 2: Stroke type versus time from symptom onset (TSO).Graphic illustration of the 102 patients with ischaemic infarcts and known TSO.

FIGURE 1 :
FIGURE 1: Vascular territories of ischaemic infarcts (n = 118).TABLE2: Order to report time and arrival to report time compared with international recommendations (American Heart Association).

TABLE 1 :
Demographics and risk profile.
†, Information on risk factors was available for only 133 of the patients.

TABLE 2 :
Order to report time and arrival to report time compared with international recommendations (American Heart Association).
ORT, order to report time; ART, arrival to report time; DTIT, door to imaging time; RT, report time.