In this retrospective study we have demonstrated the feasibility of successfully obtaining MRI scans of the head of infants without general anaesthesia using the feed-and-swaddle technique, with an overall success rate of 76.1%. We have found that for MRI brain very few scans (3/34) were unsuccessful. For MRI CPA scans the success rate was higher for younger infants, and for female patients compared to male patients. The other variables (hearing loss and time of day) did not show an association with the success rate of MRI CPA.

Various studies have already described the feed-and-swaddle technique for obtaining MRI scans of infants, with success rates for children ranging from 52% up to 100% [10,11,12,13,14,15,16, 19,20,21,22].

The reported success rates are in general higher than what we found in our study, which can largely be attributed to the type of MRI under investigation, or the age of the patients included in the studies. Fogel et al. [19], Shariat et al. [21] and Windram et al. [22] have reported success rates of 96–100%, but those studies focused specifically on cardiac or cardiovascular MRI, which might be less sensitive to motion artefacts. Success rates of 100% and 89% were reported by Golan et al. [20] and Hansen et al. [11], respectively. However, these studies did not focus on a specific type of MRI. Furthermore, in the study of Hansen et al. the majority of patients was younger than 1 week, with only a few patients aged over 12 weeks. Success rates from 79% (completely addressed the clinical question) up to 99% (at least partly addressed the clinical question) were described by Antonov et al. [12] for all types of MRI scans, but they have only included patients younger than 92 days, with the majority of included patients scanned at an age of less than 1 month. Moreover, only 1 patient was referred from the department of ENT. For full-body MRI without general anaesthesia a success rate of 94% has been reported by Gale et al. [10]. Although this was an extensive study, involving over 700 patients, most of the patients included were under 1 month of age, whereas approximately half of the patients in our study were aged 3 months or older.

Our results indicate that success rates of MRI scans of the head, particularly of MRI CPA for diagnostic purposes, cannot be directly adopted from success rates for other types of MRI scans. Imaging of anatomical structures related to hearing loss requires a high accuracy, making the scans sensitive to motion artefacts and thereby decreasing the chance of a successful scan. This assumption is supported by the difference we found in success rates between the MRI CPA scans and the MRI brain scans.

Four previous studies have investigated the feed-and-swaddle technique specifically for MRI scans related to sensorineural hearing loss (SNHL) [13,14,15,16]. Weng et al. only included patients with a corrected age of less than 13 weeks and found a success rate of 86.8% [13]. A similar success percentage (86%) was reported by Grose et al. for a relatively small patient population of 21 infants, of whom 14 were younger than 12 weeks [14]. A success percentage of only 52% was found by Liao et al., for patients with a median age of 3.2 months [16]. Ronner et al. reported a success rate of 82% for patients with a median age of 2.11 months [15].

The success rates found in our study are in general slightly lower than those presented in previous studies. When taking into consideration the type of scan and age of the infants, our results are in line with existing literature, but they provide more realistic expectations for an ENT clinic aiming to implement the feed-and-swaddle technique.

Literature addressing factors influencing the chance of a successful MRI scan is limited. We found that the type of requested scan is an important predictor for success. This might be explained by differences in required accuracy for different scans and purposes.

Another possible explanation is a difference in scan duration or noise level of the scan. Unfortunately, information on duration or noise level of specific sequences is not available, neither within our hospital database, nor in existing literature.

Specifically for MRI CPA we found that scanning at a younger age is associated with a higher chance of success. Such an effect has also been reported by Weng et al. [13]. For scans related to hearing loss, Ronner et al. and Liao et al. did not find an effect of age on success chance [15, 16], but this is possibly due to small sample sizes. Liao et al. do imply a negative effect of increasing age on success chance might be present and become significant with a larger sample size.

The effect of sex on successful scanning has not been reported in previous literature. Liao et al. did find an odds ratio of 2.05 for successful scanning of female infants compared to male infants, but they could not prove the significance of this effect.

Two factors that have been reported to reduce the success rate of scanning without general anaesthesia are prematurity and comorbidity [12, 16]. Due to the presence of only 1 premature patient in our database and the wide variety of possible comorbidities, these factors were not considered as feasible separate independent variables in our study. We were therefore unable to verify the associations between prematurity and comorbidity with chance of success.

A final factor possibly related to successful scanning is hearing loss. Infants who are less able to hear the noise of the MRI scanner, are expected to be more likely to successfully undergo an MRI without general anaesthesia. Ronner et al. indeed found an increase in success rate for infants with bilateral profound SNHL, compared to infants with other types of hearing loss. However, when comparing bilateral SNHL with unilateral SNHL, they did not find a significant effect [15]. The results from Grose et al. suggest a link between bilateral hearing loss and an increased success rate compared to unilateral hearing loss, but they did not report a statistical test [14]. A positive, but insignificant relation between the severity of hearing loss and the chance of success was also mentioned by Liao et al. [16]. Our results show the success rates for patients with profound bilateral hearing loss are slightly higher than those for patients with mild to severe bilateral hearing loss as well, but this difference was not significant. This difference cannot be seen when comparing profound bilateral hearing loss to no/unilateral hearing loss. We are therefore unable to support the assumption that (profound) bilateral hearing loss leads to an increased chance of a successful scan. Possibly, other factors surrounding the MRI, such as temperature of the scanning room or movements of the scanning table, cause these infants to wake up, despite being less able or unable to hear the scanning noises. Further research with larger sample sizes might be able to reveal an association between hearing loss and successful scanning without general anaesthesia. However, based on our results we expect that an effect of hearing loss, if it exists, would be too small to be of clinical relevance for individual patients.

In this study we have provided proof of the feasibility of MRI scanning of the head without general anaesthesia. Additionally, we identified patient factors associated with the chance of success. There is very little information available on success factors of the feed-and-swaddle technique for MRI in general and even fewer studies have specifically investigated MRI of the head. Our results give clinicians requesting MRI scans for infants the opportunity to discuss the chances of a successful scan without general anaesthesia during consultation with parents or caregivers, to give a more personal estimate of the success chance and enable shared decision making. Our results can also be used to formulate a policy for scanning infants using the feed-and-swaddle technique, which optimizes the number of MRI scans made without general anaesthesia, while minimizing the number of failed attempts. Such a policy could consist of the advice to scan as early as possible, at least during the first three months of life, as this increases the chance of a successful scan.

Another strength of our study, next to the clinical applicability, is our relatively large sample size, compared to other studies that specifically investigated success of feed-and-swaddle MRI for infants with hearing loss [13,14,15,16].

Although our sample size is relatively large compared to many previous studies, considering the many factors potentially influencing the success of the scan, the sample size would preferably be even larger. A larger sample size would potentially allow for identifying additional relevant factors, such as hearing loss. It would also allow for a more precise evaluation of the effect of age on success by categorising age into smaller categories.

Since this is a retrospective study, the observed success rates might have been influenced by prior assessments of the physicians referring for a scan. Although our centre does not use formal eligibility criteria for applying the feed-and-swaddle technique, multiple physicians have mentioned they are more inclined to attempt scanning without general anaesthesia if the infant is younger. Our data indeed shows that for older infants, scans have frequently been obtained using anaesthesia, especially for MRI of the brain. Presuming the same age-relation observed for MRI CPA, applies to MRI brain, we expect that if all these patients had undergone MRI without anaesthesia, the overall success rates would have been lower.

Another limitation related to the availability of data is the possible influence of external factors on the odds of success. During the interviews, multiple physicians mentioned that tranquillity and patience are key in successfully obtaining an MRI of children. This means the odds of success might also be affected by external factors, such as patience of the MRI technician, time allocated for repeating sequences, noisiness of the waiting area, and presence of a parent in the scanning room. Such factors are not saved in the patient file and have therefore not been included. However, since all scans were made within the same centre, it is expected that these external factors have been similar for most patients, rendering it unlikely that the associations found in this research should actually be attributed to external factors.

A final limitation of this study is the uncertainty in the age of infants. The age at time of scanning has been corrected for prematurity based on a standard of 38 weeks pregnancy, but this correction was only applied for children born prematurely. For children born a term only the actual age was considered. Since full term babies are born at 37 up to 41 weeks pregnancy and pregnancy duration is not saved in the patient file, there is an uncertainty of approximately one month in the corrected age of infants. It should be noted that previous studies have shown an association between sex of the foetus and duration of pregnancy, where male foetuses have been associated with an increase in preterm births as well as prolonged pregnancies [23, 24]. This means that taking into account the corrected age compared to a standard of 38 weeks pregnancy for all patients, both premature and full or late term, might explain the difference between male and female patients found in our study.

This study shows that using the feed-and-swaddle procedure to obtain an MRI of infants under 6 months is feasible, making it a suitable alternative to scanning under general anaesthesia. When deciding to use the feed-and-swaddle method, a referring physician should consider the type of scan, since the success rate for MRI CPA is significantly lower than for MRI brain. Specifically for MRI CPA it is important to consider the age of the patient. The chance of success for younger infants (under 3 months) is higher than for older infants (3–5 months). It might be advisable to only scan infants under 3 months using the feed-and-swaddle procedure and to use general anaesthesia for older infants, especially if scanning capacity is an issue or rapid scanning results are of great importance. Our results can also be used in shared decision-making with parents or caregivers. An example of this could be to discuss with caregivers that the chance of a failed MRI CPA scan is higher for boys, especially if they are 3 months or older.