Interventional procedure overview of phrenic nerve pacing for ventilator-dependent high cervical spinal cord injury
Closed for comments This consultation ended on at Request commenting lead permission
Evidence summary
Population and studies description
This interventional procedure overview is based on 1,647 people who had the procedure from 4 non-randomised comparative studies (Hirschfeld 2022, 2008; Romero 2012; Andersen 2017) and 1 analysis of the ADB database (Headley 2023). This is a rapid review of the literature, and a flow chart of the complete selection process is shown in figure 1. This overview presents 5 studies as the key evidence in table 2 and table 3, and lists other relevant studies in table 5.
Of the 4 non-randomised comparative studies included in the key evidence, 2 studies (Hirschfeld 2022, 2008) were done in Germany, 1 study in Spain (Romero 2012), and 1 study in Denmark (Andersen 2017). There might be an overlap in the samples between Hirschfeld (2008) and Hirschfeld (2022).
These 4 studies included a total of 305 people who were ventilator-dependent caused by SCIs (n=303) and CHS (n=2). When reported, there were 222 people with traumatic SCIs and 58 people with non-traumatic SCIs (Hirschfeld 2008, 2023; Romero 2012). At baseline, age at injury was the main notable difference between people on PNP and those on MV. Hirschfeld (2022, 2008) and Romero (2012) noted that people on PNP were statistically significantly younger than those on MV. Although Andersen (2017) reported a younger age in the PNP group than the MV group, the difference was not statistically significant. There was no statistically significant difference in ASIA classification (Hirschfeld 2008; Romero 2012), but a statistically significant difference in SCI level between groups was noted in Romero (2012). The time interval from injury to PNP implantation was mean 21 months (Romero 2012) or median 1.47 years (Hirschfeld 2022). The follow up or observational period across 4 studies ranged from 10 to 33 years.
The analysis of the ABD database (Headley 2023) included people with different indications (including SCIs) and potentially from different countries, but the exact number of people with ventilator-dependent high cervical SCIs was not reported. But it reviewed data collected over 38 years and included 1,522 people who had the Avery device implanted. Also, it particularly reported revision data and detailed the reasons for revisions (safety data). Table 2 presents study details.
Procedure technique
All 4 non-randomised comparative studies described the procedure technique but varied in detail. Two devices (Avery and Atrotech) were used.
The function of PN and diaphragm muscles was ascertained by neurophysiologic and fluoroscopic or sonographic studies (Hirschfeld 2008). This was done by applying external current to the nerve transcutaneously on the neck, while measuring diaphragm contraction with fluoroscopy or ultrasound (Andersen 2017).
When reported, both bipolar and 4‑pole electrodes were used, with the latter used more frequently. The common approach to implantation was by an open thoracotomy, with the minority using a thoracoscopic technique. The procedure was usually done bilaterally.
After the procedure, the mean conditioning duration was 47 days (Romero 2012) to 50 days (Hirschfeld 2022).
For the review of the ABD database, Headley (2023) generally described the procedure technique with a cervical or thoracic approach. The authors stated that positive identification of the PN was achieved with a disposable nerve stimulator revealing diaphragm movement. The cervical technique remained the most minimally invasive technique and could be done under local anaesthesia. But, there was a greater area of accessible PN in the chest for placement of the electrode, so more thoracic surgeons have been practicing thoracic placement (using either the open thoracostomy or the less invasive VATS technique) as opposed to cervical placement. Also, thoracic placement of the electrodes was more common in children.
Efficacy
Survival
Survival data was reported in 3 studies, with a tendency towards better survival with PNP than with MV. Hirschfeld (2022) reported that, for the whole study population of 92 people, the survival rate was 92% after 1 year, 63% after 10 years, and 61% after 20 years. A difference was in favour of PNP. People on PNP were younger than those on MV (p=0.001) at injury and increasing age at injury correlated with decreasing survival (p=0.009).
Hirschfeld (2008) found that the survival rate was 63% (20 of 32) in people on PNP and 56% (18 of 32) in people on MV over 22 years. A trend was in favour of PNP, but the difference compared with MV was not statistically significant (log-rank p=0.184).
Romero (2012) described that the survival rate was 74% (28 of 38) in people on PNP and 44% (39 of 88) in people on MV over 10 years. The length of survival was statistically significantly longer in the PNP group than the MV group (PNP, 21.78 years; MV, 8.69 years; p<0.001). Once age was adjusted, the length of survival was greater for people on PNP than those with MV (p=0.04).
Mode of ventilation or daily pacing duration
Mode of ventilation or daily pacing hours was reported in 4 studies. PNP represented the warning from MV at various levels, with 25% to 74% of people using PNP as their only mode of ventilation and the daily pacing duration between 7 and 24 hours.
Hirschfeld (2022) reported that of the 92 ventilator-dependent people, 48 people were on PNP and 44 were on MV. For the PNP group, 63% (30 of 48) of people used PNP full-time (24 hours) and 38% (18 of 48) of people used PNP intermittently (with 16 of them using MV during sleep for safety reasons). For the 48 people on PNP, 75% (36 of 48) of people continued to use a tracheal cannula, 19% (9 of 48) had their tracheostoma plugged, and 6% (3 of 48) had the tracheostoma closed.
Hirschfeld (2008) found that all 64 people (32 on PNP and 32 on MV) needed their respiratory devices during sleep. Ten people on PNP and 4 on MV used their device part time, so they could use glossopharyngeal breathing or their accessory respiratory muscles in the neck intermittently. This facilitates nursing and improves the chance to survive respiratory device failure.
Romero (2012) reported that, of the 38 people on PNP who needed permanent MV before the procedure, 27 people survived over 10 years after implantation. Of these people, 74% (20 of 27) people used PNP as their only mode of ventilation, 19% (5 of 27) used a mixed model (PNP as the main mode and MV during sleep hours), and 7% (2 of 27) moved to MV (1 had the PNP explanted because of infection and the other could no longer use it because of PN degeneration).
In the Anderson (2017) study of 14 ventilator-dependent people, 7 people had a PN pacer implanted. Their daily pacing duration was 8.5 hours in 1 person, 12 to 14 hours in 3 people and 16 hours in 3 people. This was in line with the recommendation from the study site's respiratory centre that no people should use the pacer for more than 16 hours per day to prevent any damage to the PN. The total period of pacing ranged from 1.5 years to 25.4 years.
In the analysis of the ABD database, Headley (2023) reported that, of the 111 people who responded to the survey, 57% of respondents reported that they paced for 7 to 12 hours daily, primarily while sleeping, 14% reported 13 to 15 hours daily, 13% reported 16 to 20 hours daily, and 16% used the pacer at all times. The authors also found that 5 people were pacing for 40 years (2 of these people reported as now deceased in the database), and 41 people were pacing for over 30 years (8 of these people reported as now deceased). The survey results also showed that 76% of respondents had a tracheostomy before implantation, and of these people, around 33% chose to have it removed following implantation.
Incidence of RIs
RI was recorded when the person presented with fever, leucocytosis, increased production of secretions and the doctor in charge diagnosed the reason to be RI with antimicrobial treatment being necessary. The data on the incidence of RIs or pneumonias was reported in 3 studies. There were statistically significant reductions in the incidence of RIs after implantation and fewer RIs in the PNP group than the MV group. This difference was not found in pneumonia.
Hirschfeld (2022) reported that the incidence of RIs after discharge was statistically significantly lower in people using PNP (30 people on PNP for 24 hours a day, 0.07 RIs per 100 days; 18 people on PNP intermittently, 0.08 RIs per 100 days) than people using MV (n=44, 0.2 RIs per 100 days; p=0.000).
Hirschfeld (2008) reported incidence of RIs in 3 periods: period 1 (reported as 120 days in institution before using final respiratory device); period 2 (reported as from beginning of use of final device until leave from institution for final location) and period 3 (reported as after arrival at final location, including total time of follow up, at least 1 year). There was no significant difference in the median incidence of RIs between the PNP group (n=32) and the MV group (n=32) in period 1. But, during post-implantation (period 2 and period 3) there were statistically significantly fewer RIs in the PNP group than the MV group:
Period 1: 1.43 RIs per 100 days (IQR 0.05 to 3.92) compared with 1.33 RIs per 100 days (IQR 0.89 to 2.21); p=0.888.
Period 2: 0 RI per 100 days (IQR 0 to 0.92) compared with 2.07 RIs per 100 days (IQR 1.49 to 4.19); p<0.001.
Period 3: 0 RI per 100 days (IQR 0.00 to 0.02) compared with 0.14 RIs per 100 days (IQR 0.00 to 0.31); p<0.001.
The authors also found that there were statistically significant reductions in RI incidence between periods 1 and 2, and periods 2 and 3 for each group.
Anderson (2017) found no statistically significant differences between people on PNP (n=7) and people on MV (n=7) in the number of pneumonias within the last year (median 0 compared with 0), hospitalisations for pneumonias within the last year (median 0 compared with 0), and number of daily suctions (median 3 compared with 1).
Quality of life and return to productivity
Quality of life was reported in 2 studies and various measures were used. When comparing PNP with MV, there was no statistically significant difference in quality of life across studies, except for SF‑36 social functioning improved significantly in 1 study.
Romero (2012) reported that, of the 126 people, 44 people completed the SF‑36 questionnaire (24 on PNP and 20 on MV). To avoid selection bias, 36 people with C1 to C2 level ASIA A grade of SCI were included in this analysis. The results showed that people on PNP had a statistically significantly higher score in the social functioning dimension (p=0.0002) than people on MV, but no statistically significant differences in other domains and in the total scores (all p>0.05).
Andersen (2017) did not find any significant differences in SWLS, ISCIQoLBDS (general, physical and mental domains) and SF‑36 (physical and mental domains) between the PNP group (n=7) and the MV group (n=7).
In terms of return to productivity, the data was described in 2 studies and age might play an important role in this outcome. Hirschfeld (2008) reported that of 64 people, 38 people (20 on PNP and 18 on MV) lived at home at the end of the study. Of these 38 people, 7 people on PNP and 2 on MV returned to school, 2 people on PNP but none on MV returned to work and all others retired. Anderson (2017) reported that of the 14 people, after their SCI, only 1 person on PNP completed an education. One person on MV was completing their education and another was taking some occasional educational courses.
Quality of speech
Quality of speech was evaluated in 2 studies with mixed outcomes. With pressure-controlled MV, people talk during inspiration; with PNP, people talk during expiration. Hirschfeld (2008) reported that the quality of speech was statistically significantly better in people with PNP (median score 6 [normal voice]), than people with MV (median score 3.5 [between intermittently low voice and low voice], p<0.001), but no difference between groups in the ability to talk (exact data not reported). Anderson (2017) found no statistically significant difference in the quality of speech between the PNP group and the MV group (median score 5 [intermittently normal voice] for each group).
Implant longevity
Implant longevity was indicated by years in between revision surgeries and presented in the Headley (2023) study. The authors reviewed the ABD database of 1,522 people and found that the mean longevity was 6.5 years (SD 6.2) for both cervical and thoracic approaches. When comparing 2 approaches, there was no statistically significant difference in device longevity (cervically implanted device: mean 6.4 years, SD 6.8; thoracically implanted: mean 6.4 years, SD 5.7; p=0.9382).
Safety
Mortality and causes
Mortality and its causes were described in 3 studies and the mortality rate ranged from 29% to 40% in people on PNP. Hirschfeld (2022) reported that, of 92 people, 37 people (PNP, n=19 [40%]; MV, n=18 [41%]) died over a 33-year time period. The leading cause was pneumonia (4 PNP and MV, 1 PNP used for 24 hours a day, and 10 MV). The other reasons were non-specific SCI-induced (3 MV), intestinal occlusion (2 PNP, 2 MV), decubital sepsis (1 PNP, 1 MV), urosepsis (MV), tumour and myelitis (PNP), heart failure (4 PNP and 1 MV), suicide (2 PNP and 1 MV), and seizures (1 PNP) and bleeding (1 PNP).
Hirshfeld (2008) reported that 12 people (38%) on PNP and 14 people (44%) on MV died over a 22-year time period (p=0.1023); of these people, 3 on PNP and 10 on MV died of RIs (p=0.0472).
Romero (2012) reported that the mortality rate was 29% (11 of 38) of people on PNP and 56% (49 of 88) of people on MV over 10 years. Most died from respiratory causes.
Revision and causes
Headley (2023) reported that, of 854 people who had the current version of the receiver implanted (version I-110), 20% (172 of 854) needed revision surgeries. For the electrodes, 17% (66 of 380) of cervical cases needed at least 1 revision of the electrode compared with 18% (95 of 518) of thoracic cases. Data showed that in people initially implanted cervically, 45% of electrode revisions involved moving the electrode placement to the chest compared with 10% of people whose implants were moved from the chest to the neck. The authors also reported the revision rationale for people with cervical implantation as follows:
no report or no problem found: 18%
surgical placement of implants: 14%
intermittent (loss of stimulation): 14%
insulation damage: 12%
damage to wire: 9%
calcification of anode: 8%
accidental damage (sports): 6%
accidental damage (medical treatment): 5%
infection after surgery: 5%
people who play or fidget with their subcutaneously placed receivers: 4%
patient growth: 5%
Hirschfeld (2022) reported that within 5 years after the procedure, there were 15 complications needing revisions in 13 people (27%) because of bilateral PN pacers implantation. These included falling electrode sites (n=5), haemothorax or pneumothorax (n=5), nerve loss (n=1), dislocated stimulator (n=1) and falling stimulator (n=1).
PN damage
PN damage was reported in 2 studies. Headley (2023) found that, of the 3,478 implants, PN injury caused by surgical manipulation of the nerve was reported in 6 cases over 38 years (less than 0.2%). In 5 of the 6 cases the nerve function recovered. Hirschfeld (2022) reported that, of the 92 people, the rate of nerves at risk was 4% (7.3 of 184) and 1 nerve (0.5%) was lost.
Anecdotal and theoretical adverse events
Expert advice was sought from consultants who have been nominated or ratified by their professional society or royal college. They were asked if they knew of any other adverse events for this procedure that they had heard about (anecdotal), which were not reported in the literature. They were also asked if they thought there were other adverse events that might possibly occur, even if they had never happened (theoretical).
They listed the following anecdotal or theoretical adverse events: malfunctioning of the device needing replacement, respiratory failure, bleeding, infection and injury to chest organs.
Two professional expert questionnaires for this procedure were submitted. Find full details of what the professional expert said about the procedure in the specialist advice questionnaires for this procedure.
Validity and generalisability
Of the 4 non-randomised comparative studies, 2 studies were prospective, 1 study was a retrospective analysis of prospectively collected data with a follow-up questionnaire, and 1 study was retrospective with a follow-up questionnaire in design. The sample size ranged from 23 to 92 people across these 4 studies. When considering the number of people with SCIs who used PNP, the sample size ranged from 7 to 46 people. The follow-up or observational period was between 10 and 33 years after implantation.
For the review of the ABD database, Headley (2023) included a large sample with mixed indications, but the exact number of people with ventilator-dependent high cervical SCIs was unknown. This review mainly focused on the revision aspect, so there was a lack of other outcomes of interest reported.
Across all studies, only 1 paper's authors declared their conflicts of interest (Headley 2023), while other papers did not report. None of the studies were funded by manufacturers.
Overall, the evidence suggests a tendency towards better survival with PNP than with MV. This might be due to younger age rather than the type of respiratory treatment and people on PNP were generally younger than those on MV (Hirschfeld 2022). Younger age (being more active) could also be 1 of the reasons for better return to productivity with PNP than with MV (Hirschfeld 2008).
PNP represented the weaning from MV at various levels and the longest length of pacing was over 40 years. There were statistically significant reductions in the incidence of RIs after implantation and fewer RIs in the PNP group than the MV group but not for pneumonia.
Regarding quality of life, when comparing PNP with MV, there were no statistically significant differences across studies, except for SF-36 social functioning improved significantly in 1 study. This improvement might be due to people's preference of PNP with improved portability, without tubes and with low maintenance requirements (Hirschfeld 2022; Romero 2012). However, the SF-36 questionnaire was not an ideal measurement for people with SCIs, in particular 'nominated physical functioning' (Romero 2012). Also, the quality of speech was assessed, demonstrating mixed outcomes, and no validated tool was available.
Implant longevity was measured by years in between revision surgeries, with the mean longevity being 6.5 years. The data on revision mainly came from a review of the ABD database, indicating that the rate of revision surgeries for the I-110 receiver (current version) was 20%. The rate of mortality ranged from 29% to 40% in people with PNP, with the leading cause being RIs. Other complications, such as PN damage, were rare.
In conclusion, the evidence shows improvements in outcomes. Although some improvements were limited, these limited improvements must be interpreted in the context of people with high cervical SCI who have multiple comorbidities and this procedure is to treat one component of a very complex condition. To date, no ongoing trials have been identified.
How are you taking part in this consultation?
You will not be able to change how you comment later.
You must be signed in to answer questions