Correction and response to: Grommets in HBOT patients: GA vs. LA, unanswered questions.

We thank Gibbs and Commons for their interest in our paper. There is a key difference between the datasets for Commons et al and our study. Our data set, has grouped five years of data according to the calendar year. This is different from Commons et al's study population recruited between 01 June 2009 and 31 May 2010. We feel this may explain the difference of one case between the two papers in 2010. Our data collection used the standard clinic and operating theatre databases, and we were advised that there was no searchable clinical code for grommet procedures undertaken with local anaesthetic (LA) in the outpatient clinic. The alternative, to review many hundreds of patients, was considered beyond the study's scope. Instead, the TTH Hyperbaric Medicine Unit (HMU) database was used to recruit cases and cross checked with operating theatre data. We have since re-investigated the operating theatre database to identify any additional bilateral grommet procedures during 2008 to 2012 and cross checked these with the HMU database. This has identified one to four additional patients per year in the general anaesthesia (GA) group and one additional LA patient that meet the criteria for recruitment into the study. There was one further unconfirmed patient from each of 2008 and 2010, whose charts were unavailable for this response, and have not been included in this amendment. The corrected Figure 1 reflects these changes. Despite the additional cases, the frequency spike during 2010 remains. A published audit of the number of middle ear barotrauma (MEBT) cases between 2007-2010 also reports an increased incidence of MEBT in 2009-2010 compared with previous years at our unit. Possible reasons for this may be the introduction of new technology at the unit, in the form of the digital Macro View™ otoscope during this period, leading to a possible change in clinical practice and an increased detection of MEBT, or a lower threshold for ENT referral for grommet placement. Alternatively, a 'Hawthorne effect' from the conduct of a prospective study within the TTH HMU, during 2009-2010 may be considered. With the outliers removed using ROUT's test, the significant difference in the delay time to surgery remains (LA median 1, IQR 2, range 0-5 days; GA median 7.5, IQR 6, range 0-24 days; P < 0.0001; Figure 2). The data values of 98 days and 86 days from the GA group published in our paper are corrected to six days and 12 days respectively. On review, the first individual was found to have had two HBOT courses, and it was only in the second round of HBOT that an ENT referral for grommets was made. The second individual was found to have been offered two ENT referrals after experiencing MEBT, the first was followed by the patient declining further HBOT until representing to TTH HMU four months later and receiving prophylactic grommets before recommencing HBOT; this second ENT referral date has been used in the amended data. These corrections have not been found to change the primary outcome of statistical significance between the LA and GA groups. A delay of seven days may not be considered clinically relevant in the most common cases requiring HBOT, aside from affecting patient convenience and logistics as well as hospital efficiency and resources. In emergency cases, knowledge of factors able to reduce the delay for grommet insertion is clinically relevant. In centres where a long wait for GA is the norm, LA may convey a clinically important lesser waiting time. As a retrospective study, only data documented in the patient records could be studied, and patient discomfort was not consistently recorded in the charts. We would liken this to undertaking other surgical procedures, where clinicians often do not routinely document pain scores for the benefit of retrospective research. Several studies have examined patients' tolerance of grommets under LA, finding the technique tolerable.