Keywords
Capsule endoscopy - Small bowel endoscopy - Epidemiology
Introduction
Thin and long, positioned between the stomach and the colon, and a long distance from
mouth or anus, the small bowel (SB) has long been poorly accessible to physicians.
Previously, radiological imaging was limited to SB follow-through. Since then, SB
cross-sectional imaging (computed tomography scan, magnetic resonance imaging [MRI]),
with or without enteroclysis, has been made available, allowing more precise but still
indirect investigations. At the start of the century, the advent of capsule endoscopy
(CE) allowed direct visualization of the SB in patients with suspected small bowel
bleeding (SSBB), suspected Crohn’s disease, and other conditions [1], and more recently, artificial intelligence is being used to speed up the reading
time for CE [2]. In parallel to CE, push and then device-assisted deep enteroscopy (DE) was developed;
during these invasive and labor-intensive procedures, specifically-trained operators
gain access to lesions in the SB, manipulating overtubes, balloons (single or double)
or spirals (possibly motorized) in patients under general anesthesia [3]. Given that CE is non-invasive and offers direct visualization, it is preferred
to cross-sectional imaging for diagnosis (but these can be combined), whereas indications
for device-assisted enteroscopy (DAE) are restricted to challenging diagnoses requiring
biopsies and to direct endoscopic treatments such as hemostasis, polypectomy, and
dilatations. Little is currently known about SB CE and DAE procedure volumes at national
levels, and whether these volumes would meet clinical needs in large populations.
Therefore, we aimed to describe the availability of SB endoscopy (CE and DAE) in France
from 2015 to 2021.
Patients and methods
Data source
This study was conducted using the French national health data system (Système National
des Données de Santé, SNDS), which is a database that includes comprehensive individual
medical information for 99% of the population living in France (approximately 67 million
people) since 2006 [4]. The database contains data about all outpatient procedures reimbursed by the
national health insurance system, including medication and physician visits. We also
extracted data from the French national hospital discharge database (Programme de
Médicalisation des Systèmes d’Information, PMSI) that comprises all procedures performed
during hospital stays (including one-day clinics) in both public and private institutions
(but does not include those performed in medical offices outside of institutions).
All SB CE
(French medical classification for clinical procedures, HGQD002) and DAE (anterograde
enteroscopy [HGQE003], retrograde enteroscopy [HGQE005], combined anterograde and
retrograde
route enteroscopy [HGQE001], hemostasis during anterograde enteroscopy [HGSE001 or
HGSE002],
polypectomy during anterograde enteroscopy [HGFE003 or HGFE004], hemostasis during
retrograde enteroscopy [HGSE003 or HGSE004], polypectomy during retrograde enteroscopy
[HGFE001 or HGFE002]) were extracted from these databases [5]; per-operative enteroscopy (HGQE004) was not taken into account. Dilatations of
ileal stenosis (HGAE001) were excluded from data collection because most procedures
are
performed during colonoscopy (or rectoscopy in patients with subtotal colectomy),
and more
rarely during enteroscopies. Enteroscopy-assisted endoscopic retrograde
cholangiopancreatography were also excluded.
For DAE specifically, we crosschecked data with that of enteroscope manufacturers
in France. Only data from centers equipped and/or with maintenance contracts between
January 2012 to December 2021 with Fujifilm (Tokyo, Japan) double-balloon compatible
enteroscopes (EN-450T5, EN-580T), with Olympus (Tokyo, Japan) single-balloon enteroscopes
(SIF-180), or with Olympus motorized spiral enteroscopes (PSF-1) were included in
the study. Other centers were excluded, as they possibly used codes for enteroscopy
inappropriately, and/or performed push, but not device-assisted, enteroscopy. Of note,
data were analyzed per procedure, although some patients may have combined (CE then
DAE) or repeated procedures (multiple CE and/or DAE, in the same year or over years).
We identified all procedures performed in France between January 2015 and December
2021.
The numbers of SB endoscopy centers and procedures were described by year, type of
procedure
(SB CE, diagnostic DAE, therapeutic DAE, “anterograde” or “retrograde” or “combined
anterograde and retrograde” DAE), type of practice (public vs private), and area (according
to the 13 French administrative regions in the mainland [Auvergne-Rhône-Alpes,
Bourgogne-Franche-Comté, Bretagne, Centre-Val de Loire, Corse, Grand Est, Hauts-de-France,
Ile-de-France, Normandie, Nouvelle-Aquitaine, Occitanie, Pays de la Loire,
Provence-Alpes-Côte d'Azur], and to overseas regions [Outre-Mer]). These numbers were
given
in relation to the number of inhabitants per region (according to the yearly national
census
[6]).
Ethics
Because the information used in the study was related to only procedure codes, informed
consent was not required.
Statistical analysis
Descriptive analysis (numbers and proportions) was performed during the study period
using Microsoft Excel (Microsoft, Redmond, Washington, United States).
Results
A total of 181,526 enteroscopy procedures were identified over the study period, including
151,096 specific SB CE procedures. After exclusion of centers not equipped with DAE
systems
(according to manufacturer’s data, n = 23,300), and procedures that may not require
DAE
systems (ileal dilatation during retrograde enteroscopy, n = 307; per-operative enteroscopy,
n
= 21), there were 6,802 DAE procedures included ([Fig. 1]).
Fig. 1 Flowchart. Extraction and selection of small bowel capsule endoscopy and device-assisted
enteroscopy (DAE) procedures from the French national hospital discharge database
and from the French national health data system (Système National des Données de Santé,
[5]).
Small bowel capsule endoscopy (SB CE) procedure load
Nationwide, the total number of SBCEs increased from 18,956 in 2015 to 24,183 in 2021
(+27.6%; [Fig. 2]). The majority (58.9%) of SBCEs in the 2015 to 2021 period were
performed in public institutions. The number of medical institutions in which SBCEs
were
performed varied from 757 in 2015, up to 828 in 2017, and to 796 in 2019 (+5.2%),
and then
dropped to 701 in 2020 and 710 in 2021 (mainly due to the disengagement of 90 public
institutions in 2020; Supplementary Figure 1, supplementary material). The proportion
of
private institutions offering SBCEs increased over the study period from 22.6% in
2015 to
33.0% in 2021.
Fig. 2 Number of small bowel capsule endoscopy procedures performed in France each year from
2015 to 2021 according to the type of practice (public or private).
Fig. 3 Number of device-assisted enteroscopies performed in France from 2015 to 2021, according
to the year and type of practice (public or private).
At the regional level, the total number of SBCEs performed increased between 2015
and
2021 in all regions except Centre-Val de Loire (–6.7%) and Corse (–14.7%). Most regions
experienced a drop in the number of SBCEs performed in 2020 compared to 2019, except
in
Hauts-de-France (+0.6%), in Pays de la Loire (+3.3%), and Overseas (where it markedly
increased by +11.4% in 2020 with respect to 2019, but by a small absolute number,
before
falling in 2021). Corse was the only region where SBCEs were consistently more frequently
performed in private practice than in public settings over the study period, but a
similar
pattern was observed in Pays de la Loire from 2016 to 2021 (Table 1, supplementary
material). Related to the number of inhabitants (according to 2021 census [6]), SB CE availability varied in 2021 from 142 per million overseas (and
214 per million in Centre-Val de Loire in the mainland) to 504 per million in
Hauts-de-France.
Device-assisted enteroscopy procedure load
Nationwide, the total number of DAEs decreased from 1030 in 2015 to 932 in 2021 (–9.5%),
with a reduction to 764 cases in 2020 ([Fig. 3]). The number of medical facilities where DAE were performed (23 to
26 centers) varied slightly over the study years (Table 2, supplementary material).
University hospitals provided 5,950 (87.5%) of all DAEs (n=6,802) over the study period.
The
numbers (and proportions) of DAEs coded as retrograde (only, or associated with anterograde)
DAE varied from 179 (18.8%) to 246 (22.8%) over the study years, performed in eight
to 14
centers ([Fig. 4]). The numbers (and proportions) of therapeutic DAEs (whatever
anterograde or retrograde) varied from 145 (15.6%) to 210 (20.7%) over the study period
([Fig. 5]).
Fig. 4 Number of device-assisted enteroscopies performed in France each year from 2015 to
2021 according to introduction route (anterograde, combined, or retrograde)
Fig. 5 Number of device-assisted enteroscopies performed in France each year from 2015 to
2021 according to the type of procedure (diagnostic only or therapeutic)
Fig. 6 Number of small bowel capsule endoscopy (SB CE) and device-assisted enteroscopy (DAE)
centers and procedures in the 13 mainland French administrative regions, and Overseas
(Outre-Mer), in 2015 and in 2021.
At the regional level, marked differences were noted ( [Fig. 6]). The number of centers offering DAE varied from none to five, among
the regions and the years of the study period. Two regions (Centre-Val de Loire, Corse)
did
not offer any DAE service over the study period. Two regions (Normandie, Outre-Mer)
started
offering DAE after 2015, but they were still low-volume centers in 2021 (19 and 21
procedures, respectively). Among the 10 of 14 regions offering DAE in 2015, one stopped
in
2020 (Bourgogne-Franche-Comté), six decreased (Auvergne-Rhône-Alpes –44.4%, Bretagne
–42.1%,
Grand Est –44.0%, Nouvelle-Aquitaine –53.8%, Occitanie –24.8%, Pays de la Loire –58,6%),
and
three increased the number of procedures (Hauts-de-France + 3.4%, Ile-de-France +46.1%,
Provence-Alpes-Côte d’Azur +7.3%) from 2015 to 2021 (Table 2, supplementary material).
In
2021, 22 of 24 centers (91.6%) offering DAE services were university hospitals, one
was a
private institution affiliated with a university hospital, and one was a non-affiliated
private center. All regions had DAE activity in 2019, but Nouvelle-Aquitaine and Occitanie
experienced a reduction in DAE case load in 2020. Related to the number of inhabitants
(according to 2021 census [6]), the DAE case load varied from none (in 3 regions) to 36.5 procedures
per million (in Provence-Alpes-Côte d’Azur) in 2021.
Discussion
This study shows that SBCE procedure load markedly increased from 2015 to 2021 in
France (by more than a quarter). The number of medical institutions offering CE slightly
increased from 2015 to 2019; in 2020, many public institutions disengaged (likely
because of the COVID-19 pandemic). The proportions of the 24,134 SBCEs examinations
performed in 2021 were slightly unbalanced between public (61.1%) and private settings
(38.9%). Over the same period, service in DAE had not started, had ceased or had decreased
in most (nine of 14) French administrative regions. In 2021, three regions had no
DAE service and two regions (Ile-de-France and Provence-Alpes-Côte d’Azur) accounted
for 584 (62.7%) of the 932 DAEs performed nationwide. A vast majority of DAEs (over
90%) in France were performed in university hospitals. Overall, from 2015 to 2021,
the number of SBCEs frankly increased across the country, and in both public and private
practice, whereas DAE procedure load decreased by 9.5% and was unbalanced (unequal
regional offer, mostly relying on public, university hospitals).
The present study used the French national health data system [4], compiling comprehensive data from all coded SB enteroscopies in France. However,
whereas SBCE has a very specific code that makes the data collection herein robust,
data
collection for DAE was less reliable because some procedures can be performed with
non-specific endoscopes or in patients with modified anatomy, and are assimilated
to deep
enteroscopies. In an attempt to focus on DAE, we included only enteroscopy centers
equipped
(and/or with a maintenance contract) with single or double-balloon enteroscopes 3
years before
data collection (i.e. since 2012), or more recently with motorized spiral enteroscopes,
and we
excluded perooperative enteroscopies and ileal dilatations (possibly performed with
other
endoscopes). These choices may have influenced the estimation of the number of DAEs
performed,
possibly underestimating them by excluding some procedures but possibly overestimating
them as
some procedures with colonoscopes or push enteroscopes in centers equipped with DAE
systems
may have been coded as enteroscopy; we can only acknowledge a potential bias here.
The small
(roughly 20%) proportion of DAEs listed as “therapeutic” is very surprising (when
the
procedure is mainly indicated for treatment), which also questions the validity of
the data
collected. We can only hypothesize that the specific codes for hemostasis and resections
associated with DAE were overlooked or misused by physicians. However, the external
validity
of the data collected on the overall number of DAEs is consolidated by the 25.0% reduction
in
2020 (most likely due to the COVID-19 pandemic) compared to 2019, which is similar
to that
observed for colonoscopy in France (roughly 250,000 fewer than the usual 1,400,000
yearly
colonoscopies in France, –17.8%) [7]
[8].
Previous data on SB endoscopy volumes are scarce. The United Kingdom and France have
a
similar number of inhabitants (65 and 66 million, respectively, in 2015), making comparison
worthwhile. In 2012, McAlindon et al. published the results of a web-based survey
of SB
practice throughout the UK [9]. Of 334 UK sites with gastroenterology units identified in the British Society of
Gastroenterology database, 187 were hospital-based, of which 84 (45%) offered SBCE
services.
Non-hospital-based data (if any) were not provided in this study. While procedure
load had
increased over the years, hospital-based cases had reached 8,430 SBCE procedures in
2010, all
over the country in the UK (with unequal distribution according to regions, as observed
in
France). As for volume, in direct comparison (likely increased in both countries over
time),
our first record was made 5 years later with 10,933 SB capsule procedures performed
in 586
public centers (and 8,023 in 151 private centers) in France in 2015. In 2010, 15 UK
centers
offered spiral and/or double-balloon (to be compared to 24 in France in 2015). An
Italian
nationwide web-based survey on quality performance measures provides data from 2014
to 2018
(while the Italian population was 61 million in 2015) [10]. Eighty-four centers from 17 (of 20) Italian administrative regions provided workload
data, but it is not known how many centers did not participate. From these partial
data, we
can merely conclude that the number of centers and procedures increased in Italy up
to at
least 84 and 3,899 in 2018, respectively, (to be compared to 789 centers, and 22,719
procedures in France in 2018). No data on DAE were available from the Italian initiative.
Overall, web-based, self-reported procedure loads estimated from quality surveys in
other
countries are incomplete, making comparisons poorly appropriate. Conversely, the main
strength
of our study is to provide comprehensive nationwide data on SB endoscopy procedure
loads, over
a long period of time, based on the French national health data system for outpatients,
and on
the French national hospital discharge database for inpatients (including 1-day clinics),
for
both public and private institutions. We must acknowledge, however, that 1% of the
French
population is not covered by the SNDS, that medical offices outside of institutions
are not
covered by the PMSI (but capsules and DAE are rarely performed in this setting), and
that data
on intraoperative procedures were not available. Because these limitations and volumes
are
minimal, we believe that the big picture provided by our data is valid. In addition,
we were
not able to access individual data to study the distribution of indications for SBCEs
and DAEs
over time, across regions and according to European Society for Gastrointestinal Endoscopy
(ESGE) recommendations.
SBCE service is in essence easy to set up for gastroenterologists, and quick and easy
to
perform for patients. The present study found that the technology has spread all over
the
country, in both public and private practice, since the technology was launched in
2000, and
the momentum persisted until 2019. Despite the fact that numerous public institutions
stopped
performing SBCE in 2020 (and did not resume in 2021), the procedure was less impacted
by the
COVID-19 pandemic (12.7% decrease in 2020, compared to 2019) than other, more invasive
endoscopic procedures (suggested to be –17.8% for colonoscopy) [7]. SBCE’s non-invasive nature has been used in other settings when anaesthesiologic
resources were stretched during the COVID-19 pandemic [11]
[12]. This is in contrast to the results for DAE, which seem to be the exact opposite
of
CE, including the frequency of this complex and costly procedure, use of fragile hardware,
a
frequent requirement for maintenance fees, extensive training for physicians, and
a lot of
resources for long and risky procedures [13] under general anesthesia, DAE procedure load dropped by 25% in 2020, most likely
due
to the lack of resources during the COVID-19 pandemic.
The most striking figures in this study are first, the vast difference (22-fold) in
the
number of SBCE procedures (n = 151,096) compared to the number of DAE procedures (n
= 6,802),
and second, the decoupling of the changes over time in SBCE (+27.6%) and DAE (–9.5%)
case
loads, in France over the study period. In addition, SBCE is performed throughout
the
territory, in tens of public and private institutions on the mainland, whereas DAE
procedures
were concentrated almost exclusively (> 90%) in university hospitals, mostly (62.7%)
in the
two regions where medical resources are dense (Ile-de-France and Provence-Alpes-Côte
d’Azur
were the only regions in France with 200 specialists or more for 100,000 inhabitants
in 2021)
[14]. One may hypothesize that the selection of patients for SBCE is getting loose (thus
explaining the increasing number of CEs performed, nationwide), whereas the selection
for DAE
remains strict. However, considering a worst-case scenario in which half (12,000)
of the
24,000 SBCEs in France in 2021 were indicated for SSBB (where the literature suggests
that it
is up to two-thirds of indications [1]
[15]), with a 33% diagnostic yield only, whereas the ESGE suggests a 55% to
62% yield for this indication [16]), one should speculate that there are at least 4,000 indications for
dedicated therapeutic endoscopy, i.e. mostly deep enteroscopies. Even though only
half of
cases would require DAE (because some lesions may be accessible to push enteroscopy
with a
colonoscope, for instance), one can expect 2,000 procedures every year (and this scenario
does
not even take into account other indications, with or without need for prior SBCE).
Overall,
with roughly 1,000 procedures performed every year, the number of DAEs performed in
France
seems far to meet the current need. The expert team from Sheffield (UK) suggests that
the
needed ratio between SBCE and DAE volumes is around 12 to 1 [17]. Even taking into account some likely bias regarding misused or overlooked codes
for
DAE, whereas the unique code for SBCE can be considered robust and reliable, the offer
of SB
endoscopies seems very unbalanced. This experience is shared by the Italian RAVE study
group.
The investigators noted that fewer than 40% of centers offering SBCE services were
performing
DAE procedures, this with high heterogeneity according to regions (as seen in France)
[10]. Indeed, in France, most DAE operators approached with the numbers found in our
study
confirmed that DAE is competing with many other (possibly more profitable) procedures,
and
that they are often facing endoscope shortage and increased costs due to breakage
or
preventive maintenance, and that it is really challenging to meet the current demand
for DAE
procedures (unpublished data). These findings raise questions about the rational use
of DAE,
with possible solutions to increase the current offer, aiming to approach SB diseases
better
and faster when optical diagnosis and sampling are necessary (suspicious tumors, ulcerations
or stenosis for example), to avoid surgery (including perioperative enteroscopy) when
a
non-invasive alternative option exists (SSBB and polyps, noticeably), and to reduce
delays and
costs of DAE (wherever there is a strong indication for it but it is not available
at the
institution or in the vicinity). There was a hope that spiral motorized enteroscopy,
made
commercially available in France in 2022, would change the paradigm because this technology
offered new possibilities (set up and procedure time, length of SB explored) [18]. However, it was withdrawn worldwide in mid-2023 due to safety concerns. An option
would now be to concentrate DAE procedures in a limited number of expert, fully-equipped
centers. This would call for a significant change in the current French healthcare
organization and for incentives for such centers to be developed. In addition, the
recent,
strong recommendation, based on high-quality evidence, from the ESGE to perform DAE
within the
first 48 to 72 hours after an overt SSBB favors expanding – rather than shrinking
– DAE
availability all over the French territory.
The current limited offer of DAE care in France also calls into question the availability
of DAE training (and therefore, future care). The ESGE states that training in SBCE
should be
structured with a minimum of 30 SBCE readings, but training in DAE requires first
competency
in bidirectional endoscopy and in SBCE reading, then a minimum of 75 procedures, including
35
retrograde DAEs, with at least 50 % therapeutic procedures [19]. In an environment in which DAE procedures are scarce and not available nationwide,
it
is even more challenging to ensure competency in DAE.
Conclusions
In conclusion, SB endoscopy procedure load in France from 2015 to 2021 is marked by
unbalanced offer in CE (nationwide coverage, with increasing volumes) and in DAE (decreasing
or stopping in many regions, and absent for years in others). French endoscopy opinion
leaders, medical societies and health administration should strongly consider the
causes and consequences of DAE shortcomings in terms of healthcare as the highest
priority, but also regarding training and research.
