Skip to main content

Comparison of outcomes of neurosurgical operations performed before and during the COVID-19 pandemic: a matched cohort study

Cite this dataset

Toman, Emma (2021). Comparison of outcomes of neurosurgical operations performed before and during the COVID-19 pandemic: a matched cohort study [Dataset]. Dryad.


To determine how the first wave of the COVID-19 pandemic affected outcomes for all operatively managed neurosurgical patients, not only those positive for SARS-CoV-2.

Matched cohort (pairwise method).

A single tertiary neurosurgical referral centre at a large UK Major Trauma Centre.

During the first COVID-19 wave, 231 neurosurgical cases were performed. These cases were matched to cases from 2019. Cases were matched for age (±10 years), primary pathology and surgical procedure. Cases were excluded from analysis if either the age could not be matched to within 10 years, or the primary pathology or procedure was too unique. After exclusions, 191 cases were included in final analysis

Outcome measures
Primary outcomes were 30-day mortality and postoperative pulmonary complications. Secondary outcomes included Glasgow Outcome Score (GOS) on discharge, length of stay (LoS), operative and anaesthetic times and grade of primary surgeon. An exploratory outcome was the SARS-CoV-2 status of patients.

There was no significant difference between the pandemic and matched cohorts in 30-day mortality, pulmonary complications, discharge GOS, LoS, operative or anaesthetic times. There was a significant difference in the variation of grade of primary surgeon. Only 2.2% (n=5) of patients had a SARS-CoV-2 positive swab.

During the first UK wave of the COVID-19 pandemic, the mortality, morbidity and functional outcomes of operatively managed neurosurgical patients at University Hospitals Birmingham were not significantly affected compared with normal practice. The grade of primary surgeon was significantly more senior and adds to the growing body of evidence that demonstrates how the pandemic has negatively impacted UK surgical training. Mixing COVID-19 positive, unknown and negative cases did not significantly impact on outcomes and indicates that further research is required to support the implementation of evidence-based surgical pathways, such as COVID-light sites, throughout the next stage of the pandemic.


Study design
The study was undertaken at the Queen Elizabeth Hospital Birmingham (QEHB) as a single-centre, matched cohort study. The QEHB is the largest hospital within the University Hospitals Birmingham (UHB) Foundation Trust, is one of three regional adult neurosurgical units in the West Midlands and is responsible for a mixed urban and rural population.

Operations performed during the initial stage of the pandemic (19th March 2020 to 1st June 2020) were matched to pre-pandemic cases (11th March 2019 to 11th September 2019). The pandemic time period was chosen as this was the interval in which elective neurosurgery cases at UHB were cancelled and will be referred to as the “initial pandemic phase” from here onwards. Cases were matched in a pairwise fashion for age at time of procedure (+/- 10 years), primary pathology and procedure. Primary pathology included severity where appropriate; for example, poor grade subarachnoid haemorrhage (SAH) was matched with another SAH of equivalent poor grade. Procedure was also matched to site where significant; for example a left-sided frontal craniotomy was matched with another left-sided frontal craniotomy rather than a right-sided. Ward-based procedures such as lumbar drain and ICP bolt insertion were not included. Cases were deemed unmatchable and were excluded from analysis if either the age could not be matched to within 10 years or the primary pathology or procedure was too unique.

Cases were identified by performing a search of the Galaxy electronic operating theatre system for all operations carried out under a neurosurgical consultant within the chosen time periods. Galaxy and the Prescribing Information and Communications System (PICS) were then used to collate relevant variables.

Demographic data included age and gender. Operative data included name of procedure, grade of primary surgeon and operative time points. Grade of surgeon was defined as junior Senior Registrar (SpR)(ST3-5 or equivalent), senior SpR (ST5-8 or equivalent), fellow or consultant. Pre-operative anaesthetic time was defined as time “into anaesthetic room” to “knife to skin”, operative time was defined as “knife to skin” to “skin closure” and post-operative anaesthetic time was defined as “skin closure” to time “into theatre recovery”.

Clinical data included length of hospital stay (LoS) in days, discharge Glasgow Outcome Score (GOS), (10) 30-day mortality, post-operative pulmonary complications and COVID-19 swab status. Swabs were deemed to be positive if SARS-Cov-2 RNA was detected by the standard RT-PCR laboratory test. 


  1. 30-day mortality
  2. Post-operative pulmonary complication

Primary outcomes were designed to be comparable to the COVIDSurg study.  ADDIN EN.CITE.DATA (8) For 30-day mortality, day of surgery was treated as day 0. Post-operative pulmonary complication was defined as pneumonia, acute respiratory distress syndrome (ARDS), or unexpected postoperative ventilation (any episode of non-invasive ventilation, invasive ventilation, or extracorporeal membrane oxygenation after initial extubation after surgery; or patient could not be extubated as planned after surgery).


  1. Discharge GOS (1-5)
  2. Length of stay in hospital (days)
  3. Grade of primary surgeon
  4. Pre-operative and post-operative anaesthetic time (minutes)
  5. Operative time (minutes)

The Glasgow Outcome Score is an objective, functional outcome score from 1-5. (11) Favourable GOS was defined as a score of 4 (moderate disability) or 5 (good recovery). Unfavourable GOS was defined as a score of 1 (dead), 2 (vegetative state) or 3 (severe disability).


COVID-19 status both pre- and post-operatively and the clinical course of positive patients.

Statistical methods
The study was conducted and reported in line with STROBE guidelines for cohort studies. (12) Continuous data were tested for normal distribution prior to analysis. Parametric data are presented as mean with 95% CI, and the unpaired t test used to detect differences between groups. Non-parametric data are presented as median with IQR, and Mann-Whitney used to detect differences between groups. The X2 test was used for categorical data. Missing data points were removed from final analyses. Clinical significance was defined as p<0.05.

Analyses were performed using GraphPad Prism software version 8.4.3 for Windows.