Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

bs_bs_banner

doi:10.1111/jgh.15483

GASTROENTEROLOGY

Cost-effectiveness analysis of sequential fecal microbiota


transplantation for fulminant Clostridioides difficile infection
Sanchit Gupta,*,† Jinyi Zhu,‡ Thomas R McCarty,*,† Jordan Pruce,* Zain Kassam,§ Colleen Kelly,¶,**
††
Monika Fischer and Jessica R Allegretti*,†
*Division of Gastroenterology, Hepatology, and Endoscopy, Brigham and Women’s Hospital, †Harvard Medical School, ‡Center for Health Decision Science,
Harvard School of Public Health, Boston, §Finch Therapeutics, Somerville, Massachusetts, ¶Women’s Medicine Collaborative, Lifespan, **Alpert Medical
School, Brown University, Providence, Rhode Island, and ††Division of Gastroenterology, Indiana University, Indianapolis, Indiana, USA

Key words
Abstract
Clostridioides difficile, Clostridioides difficile
infection, Cost–benefit, Cost-effectiveness Background and Aim: Fulminant Clostridioides difficile infections (FCDI) account for
analysis, Fecal microbiota transplantation, 8% of cases and substantial healthcare burden. Fecal microbiota transplantation is
Intestinal microbiota transfer. recommended for recurrent CDI, but emerging data support use for FCDI. We aimed to
assess the cost-effectiveness of a sequential fecal microbiota transplantation (sFMT)
Accepted for publication 3 February 2021. protocol for FCDI compared with current standard therapy.
Methods: A Markov model simulated patients with FCDI in a 1-year time horizon. The
Correspondence treatment algorithm for up to three sFMTs, clinical probabilities, and direct costs were used
Dr Jessica R Allegretti, 850 Boylston Street, from published sources. Outcomes were quality-adjusted life years (QALYs) and costs. The
Suite 201, Chestnut Hill, MA 02467, USA. healthcare sector perspective was used with a willingness-to-pay threshold of $100 000 per
Email: jallegretti@bwh.harvard.edu QALY.
Results: Sequential fecal microbiota transplantation (FMT) for FCDI was associated with
Declaration of conflict of interest: J. R. A. is a
lower overall cost ($28 309 vs $33 980) and higher QALY (0.765 vs 0.686) compared with
scientific advisor to OpenBiome and a
standard therapy. sFMT is cost-effective in 100% of iterations. sFMT remained
consultant for Finch Therapeutics. Z. K. is an
cost-effective at cure rates > 44.8% for the first FMT and at stool preparation cost < $6944
employee and shareholder of Finch
per instillation. We find a wide range of efficacies for the first versus second FMT at which
Therapeutics. M. F. and C. K. are scientific
sFMT is still preferred. Value of information analysis estimates the expected value of
advisors to OpenBiome. S. G., J. Z., T. R. M.,
and J. P. have no conflicts of interest.
perfect information to be low at $1.89 per person, quantified with net monetary benefit.
Financial support: S. G. was supported by NIH Conclusions: An sFMT strategy strongly dominates standard therapy, with lower cost and
Training Grant National Institute of Diabetes higher QALY. Sensitivity analysis demonstrates benefit even if FMT cure rates are lower
and Digestive and Kidney Diseases (NIDDK) than expected and when multiple FMTs are required. FMT material in 2020 was priced
T32DK007533-35. at $1695 per treatment but remains cost-effective at a much higher cost.

Introduction Various data indicate that FMT protocols, including


multiple-infusion and pseudomembrane-guided protocols, are
Clostridioides difficile infection (CDI) is a major contributor to effective for FCDI with decreased colectomy and mortality
healthcare-associated infections in the USA.1 Up to 8% of patients rates.10–16 This suggests that FMT may be considered to treat
progress to fulminant CDI (FCDI), previously termed patients with FCDI. Cost of material from a universal stool bank
severe-complicated CDI, with a 35% mortality rate.2,3 These pa- has increased over time, raising concerns regarding the
tients are characterized by hypotension or shock, ileus, megacolon, cost-effectiveness of FMT.7 While FMTs can be lifesaving, to
admission to intensive care unit, fever > 38.5 °C, mental status our knowledge, there have been no studies examining the
changes, serum lactate > 2.2 mmol/L, leukocytosis > 35 000 μL, cost-effectiveness of this approach in FCDI.
or signs of end-stage organ failure.4,5 Management of CDI, includ- Accordingly, we assessed the cost-effectiveness of a sequential
ing patients with FCDI, is estimated to require 2.4 million days of FMT (sFMT) protocol for FCDI compared with standard therapy
inpatient stay and cost $6.3bn annually.6 using oral vancomycin and intravenous metronidazole, with
Approved treatment options for FCDI are limited to vancomycin colectomy for patients not responding to antibiotics.
by mouth and/or by rectum, intravenous metronidazole, and possi-
ble colectomy.4 Fecal microbiota transplantation (FMT) is guide-
line recommended only for treating multiple recurrences of CDI Methods
and allowed under Food and Drug Administration enforcement
discretion as an investigational therapy.4 Multiple analyses have Design. We simulated a hypothetical cohort of adult patients
demonstrated cost-effectiveness of this approach in the first and with median age 65 with FCDI using a Markov model. Patients
second recurrence of CDI.7–9 meeting the 2018 Infectious Diseases Society of America/Society

Journal of Gastroenterology and Hepatology •• (2021) ••–•• 1


© 2021 Journal of Gastroenterology and Hepatology Foundation and John Wiley & Sons Australia, Ltd
Cost-effectiveness of fecal transplant S Gupta et al.

for Healthcare Epidemiology of America and 2013 American 8 weeks. For this analysis, we did not assess risk of new incident
College of Gastroenterology criteria for fulminant/severe- cases or recurrent CDI after this 8-week period.
complicated CDI were assessed by two strategies: standard of care
with medical therapy and/or colectomy, and an sFMT protocol for
those not responding to antibiotics. Model parameters. Data regarding clinical outcomes, costs,
We utilized TREEAGE PRO 2019 (TreeAge Software, Inc., and utilities were used from published sources (Table 1). Probabil-
Williamstown, MA) and a healthcare sector perspective. We com- ities of FMT efficacy and each sFMT required based on symptoms
pared direct costs and quality-adjusted life years (QALYs) of these or pseudomembranes were calculated for patients with FCDI from
strategies over a 1-year time horizon. A secondary analysis was data extracted from Fisher et al. after excluding patients with se-
conducted over a lifetime horizon, using a half-cycle correction, vere CDI.10 Perforation risk due to diagnostic colonoscopy was
3% discount rate, and average projected annual health costs. included.21
A willingness-to-pay threshold of $100 000 was considered Published medication costs were used from the
cost-effective. We assessed the impact of sFMT on costs, effective- Micromedex/RED BOOK database with a 40% discount to the av-
ness, and cost-effectiveness versus standard therapy for FCDI and erage wholesale price as per convention.24,35 Costs per day of
validated our model by comparing predicted colectomy rates and treatment of oral vancomycin 500 mg four times per day and intra-
mortality in standard therapy and sFMT strategies with observed venous metronidazole 500 mg every 8 h were calculated. Antibi-
rates in recent CDI cohorts.11,16,17 otics were administered for 14 days in standard therapy and
3-day courses for sFMT.
Hospitalization cost per day was calculated from economic data
Model structure. Figure 1 shows a state-transition diagram related to CDI and length of stay estimates by weighted averages
of the model. In the standard therapy treatment algorithm, patients from multiple sources.8,10,14,16,18,27 Length of stay for patients re-
were treated with oral vancomycin 500 mg four times per day and ceiving antibiotics has ranged from 12 to 32 days, while length of
intravenous metronidazole 1500 mg/day for 14 days. Patients with stay for patients undergoing FMT has ranged from 9 to 14 days.
refractory disease underwent colectomy. The treatment algorithm We used a conservative estimate of a 14-day hospitalization for
for up to three sFMTs in addition to antibiotics was adapted from both standard therapy and sFMT, so that expense due to hospital-
a published protocol.10 Patients without response after 3 days of ization duration would not skew the analysis. Length of stay was
antibiotics underwent FMT via colonoscopy with evaluation for multiplied by cost per day of hospitalization. Possible costs attrib-
pseudomembranous colitis. If symptoms persisted after an utable to intensive care utilization were not included. Cost of
additional 3 days of medical treatment, a second FMT was colectomy was added separately in case of that event.
performed. For patients without improvement after a third 3-day Data regarding cost of inpatient procedures including colonos-
course of antibiotics and evidence of pseudomembranous colitis, copy are limited, and despite requirements for posting prices via
a third FMT was performed. Patients without pseudomembranes the hospital chargemaster, these are unlikely reliable estimates.36
on initial FMT and without improvement underwent a second We extrapolated national data for diagnostic colonoscopy cost
FMT; if there was no improvement, they underwent colectomy. from the Centers for Medicare and Medicaid Services using Cur-
Patients without response after a third FMT and an additional rent Procedural Terminology code 45378 to more appropriately es-
3 days of antibiotics underwent colectomy. If an FMT via colonos- timate cost. Cost of FMT stool product was taken from
copy was complicated by bowel perforation, the patient underwent OpenBiome, the most commonly used stool bank with an
colectomy. Our outcomes included cure, colectomy, or death after established safety screening process, and included the 2020 cost

Figure 1 State-transition diagram of a patient with fulminant Clostridiodies difficile infection.

2 Journal of Gastroenterology and Hepatology •• (2021) ••–••


© 2021 Journal of Gastroenterology and Hepatology Foundation and John Wiley & Sons Australia, Ltd
S Gupta et al. Cost-effectiveness of fecal transplant

Table 1 Parameter probabilities, costs, and utilities

Parameter Mean Range (95% CI) Distribution Source

Clinical probabilities
18–20
Standard therapy cure 0.623 0.589–0.654 Beta Zilberberg et al., Rokas et al., and Wang et al.
10
Pseudomembranes at FMT #1 0.568 0.408–0.721 Beta Fischer et al.
10
Pseudomembranes at FMT #2 0.235 0.072–0.456 Beta Fischer et al.
10
FMT #2 if no pseudomembranes 0.143 0.032–0.317 Beta Fischer et al.
10
Cure after FMT #1 0.810 0.621–0.943 Beta Fischer et al.
10
Cure after FMT #2 0.770 0.516–0.945 Beta Fischer et al.
10
Cure after FMT #3 0.750 0.292–0.992 Beta Fischer et al.
21
Bowel perforation 0.001 0.00003–0.004 Beta Panteris et al.
10
Death after FMT #1 0.048 0.001–0.168 Beta Fischer et al.
10
Death after FMT #2 0.077 0.002–0.265 Beta Fischer et al.
10
Death after FMT #3 0.250 0.008–0.701 Beta Fischer et al.
22,23
Colectomy survival 0.644 0.610–0.677 Beta Peprah et al. and Dallas et al.
Costs (2019 USD)
24
Vancomycin for 14 days $40.72 $5.42–$43.43 Gamma Micromedex
24
Metronidazole for 14 days $49.68 $9.94–$79.50 Gamma Micromedex
24
Vancomycin for 3 days $15.17 $10.16–$21.16 Gamma Micromedex
24
Metronidazole for 3 days $26.61 $7.99–$56.23 Gamma Micromedex
25,26
FMT via colonoscopy $2339 $1754–$2923 Gamma OpenBiome and CMS
Hospitalization for 14 days $23 784 $20 297–$28 625 Gamma Rajasingham et al., Fischer et al., Ianiro et al.,
8,10,14,27
and McGlone et al.
Hospitalization for 20 days $33 978 $28 995–$40 892 Gamma Rajasingham et al., Ianiro et al., Cheng et al.,
8,14,16,18,27
Zilberberg et al., and McGlone et al.
27
Colectomy $39 047 $31 768–$47 065 Gamma McGlone et al.
Utilities
28,29
Healthy/well (median age 65) 0.88 0.86–0.92 Triangular Tengs and Wallace and Fryback et al.
9,30,31
Fulminant CDI 0.42 0.30–0.57 Triangular Varier et al., Wilcox et al., and Slobogean et al.
Post-colectomy 0.60 0.54–0.79 Triangular Hayes and Hansen, Brown et al.,
32–34
and Punekar and Hawkins
Death 0

CDI, Clostridioides difficile infection; CI, confidence interval; FMT, fecal microbiota transplantation.

adjustment.25,37 To account for inflation, all costs were converted We conducted a probabilistic sensitivity analysis (PSA) to as-
to 2019 US dollars using the consumer price index inflation calcu- sess for input uncertainty with 10 000 iterations of a
lator. Beta and gamma distributions were calculated using PARAM- second-order Monte Carlo simulation. We evaluated probability
ETER SOLVER version 3.0 (MD Anderson Cancer Center, Houston, distributions of parameter inputs and their impact on model results.
TX). For sensitivity analysis of costs of FMT via colonoscopy and Parameter distributions were calculated and PSA conducted using
colectomy, putative ranges of 25% below and above mean values a beta distribution for clinical probabilities, gamma distribution for
were calculated. Utility weights to calculate QALYs were obtained costs, and triangular distribution for utilities.
from published sources. Patients alive at the end of the 1-year time Lastly, we conducted a value of information analysis using PSA
horizon were assumed to carry average remaining lifetime mortal- for uncertainty distributions and the Sheffield Accelerated Value of
ity risk and healthcare costs.38,39 Information to evaluate the value of future research into this
question.40 The estimates describe the expected value of perfect
information (EVPI) and value of removing decision uncertainty.

Sensitivity analysis. Multiple sensitivity analyses were per-


formed. A one-way sensitivity analysis identified the threshold at Results
which cure rate of the first FMT is cost-effective. We further
assessed the FMT efficacy range with a two-way sensitivity analy- Model validation. Our simulation predicted that patients
sis to evaluate cure rates of the first and second FMTs that would with FCDI who underwent standard treatment had a colectomy
be cost-effective. Additionally, due to increases in cost of FMT rate of 15.1% and mortality rate of 15.3%, similar to observed
product, we assessed the threshold at which FMT product may colectomy rates of 15.7% and mortality rates of 14.8–
be prohibitively expensive. A secondary, exploratory threshold 21.3%.16,17 Patients with FCDI who underwent sFMT were pre-
analysis for FMT product cost was conducted with adjusted total dicted to have a colectomy rate of 2.0% and mortality rate 8.5%,
hospitalization costs using length of stay of 14 days for sFMT similar to observed colectomy rates after institution of an FMT
and 20 days for standard therapy. program of 0.7–5.5% and mortality rate 9.1%.11,16

Journal of Gastroenterology and Hepatology •• (2021) ••–•• 3


© 2021 Journal of Gastroenterology and Hepatology Foundation and John Wiley & Sons Australia, Ltd
Cost-effectiveness of fecal transplant S Gupta et al.

Table 2 Results

Base-case analysis
Treatment Cost QALY ICER
Sequential FMT 28 309 0.765 —
Standard therapy 33 980 0.686 Dominated
Probabilistic sensitivity analysis
Treatment Cost (95% CI) QALY (95% CI) ICER
Sequential FMT 28 402 (23 875–33 477) 0.765 (0.726–0.796) —
Standard therapy 33 960 (28 968–39 322) 0.686 (0.649–0.721) Dominated
One-way sensitivity analysis
1st FMT cure rate > 44.8%
Stool preparation cost < $6944
Stool preparation cost (14- vs 20-day hospitalization) < $15 601
Value of information analysis
Expected value of perfect information $1.89 per person
Value of removing decision uncertainty $942 978
Lifetime horizon analysis
Treatment Cost QALY ICER
Sequential FMT 223 956 16.77 4983
Standard therapy 214 481 14.87 —

CI, confidence interval; FMT, fecal microbiota transplantation; ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life year.

Figure 2 Net monetary benefit (NMB) evaluates the health benefits compared with incremental cost. Sequential fecal microbiota transplantation is
favored at each willingness-to-pay threshold. , sequential fecal microbiota transplantation; , standard treatment. [Color figure can be viewed at
wileyonlinelibrary.com]

4 Journal of Gastroenterology and Hepatology •• (2021) ••–••


© 2021 Journal of Gastroenterology and Hepatology Foundation and John Wiley & Sons Australia, Ltd
S Gupta et al. Cost-effectiveness of fecal transplant

Model output. Model output is described in Table 2. The sensitivity analysis for efficacy of the first two FMTs demonstrates
base-case analysis demonstrates that sFMT for FCDI was associ- a wide range of varying efficacies at which sFMT will be success-
ated with higher QALY (0.765 vs 0.686) and lower overall cost ful (Fig. 3).
($28 309 vs $33 980) when compared with standard treatment in The cost-effectiveness acceptability curve from the PSA demon-
a 1-year time horizon. The net monetary benefit at $100 000 per strates that sFMT is the preferred strategy in 100% of iterations up
QALY is $48 144 for FMT versus $34 602 for standard therapy, to and far beyond the willingness-to-pay threshold of $100 000
and sFMT is favored at all willingness-to-pay thresholds (Fig. 2). (Fig. 4). The probability FMT that is cost saving is 0.994. With un-
For the secondary analysis over a lifetime horizon, sFMT has a certainty distributions, we find that QALYs gained with sFMT
higher QALY (16.77 vs 14.87) with higher cost ($223 956 vs (0.765, 95% confidence interval [CI] 0.726–0.796) are higher than
$214 481) compared with standard therapy, with an incremental standard therapy (0.686, 95% CI 0.649–0.721) and cost is lower
cost-effectiveness ratio of $4983 per QALY, far below the with sFMT ($28 402, 95% CI $23 875–$33 477) than standard
willingness-to-pay threshold of $100 000 per QALY. therapy ($33 960, 95% CI $28 968–$39 322). Despite some over-
lap in CIs due to parameter uncertainties in these 10 000 iterations,
we see that sFMT overall maintains a higher QALY and lower cost
Sensitivity analysis. One-way sensitivity analysis demon- (Fig. 5).
strates that FMT remained cost-effective at cure rates > 44.8%
for the first FMT and at stool preparation cost < $6944 using
14-day hospitalization (Table 2). In the secondary threshold analy- Value of information analysis. The EVPI is $1.89 per
sis, when a standard treatment hospitalization estimated at 20 days person, equivalent to 0.00002 QALYs per person. Assuming that
while an sFMT hospitalization estimated at 14 days, FMT is the annual number of cases of FCDI in the USA is 25 0001,2 and
cost-effective at stool preparation cost < $15 601. Stool prepara- that FMT may be used for the next 20 years, the overall expected
tion costs do not include colonoscopy cost. The two-way value of removing decision uncertainty would be $942 978. This is

Figure 3 Two-way sensitivity analysis of cure rates for the first versus second fecal microbiota transplantation (FMT) for fulminant Clostridioides dif-
ficile infection. , sequential FMT; , standard treatment. [Color figure can be viewed at wileyonlinelibrary.com]

Journal of Gastroenterology and Hepatology •• (2021) ••–•• 5


© 2021 Journal of Gastroenterology and Hepatology Foundation and John Wiley & Sons Australia, Ltd
Cost-effectiveness of fecal transplant S Gupta et al.

Figure 4 Cost-effectiveness (CE) acceptability curve for all iterations of a sequential fecal microbiota transplantation versus standard therapy ap-
proach to treat fulminant Clostridiodes difficile infection. , sequential fecal microbiota transplantation; , standard treatment. [Color figure
can be viewed at wileyonlinelibrary.com]

the cost of removing all decision uncertainty regarding which to increased efficacy of FMT and lower likelihood of requiring
strategy is cost-effective (Table 2). colectomy than with medical therapy.
Accounting for parameter uncertainties using the PSA and sen-
sitivity analyses confirms our findings. SFMT is favored even if
FMT cure rates are as low as 44.8% for the first FMT; actual cure
Discussion rates are much higher.10,13,16 Similarly, for those who need two
We provide the first analysis evaluating the cost-effectiveness of FMTs, Figure 3 demonstrates wide ranges of cure rates over which
FMT for FCDI. We find that sFMT for FCDI is more effective sFMT is still a cost-effective approach. This is supported by the
and cost saving compared with standard treatment in a 1-year time superiority of an sFMT protocol in direct comparison with a single
horizon, as sFMT strongly dominates standard therapy with lower FMT for fulminant disease.13 FMT material from a universal stool
cost and higher QALY. Assessed over a lifetime horizon, sFMT in- bank with established donor screening protocols was priced at
creases QALY with a slightly higher cost, likely due to prolonged $1595 per treatment in 2019 and increased in price to $1695 per
life expectancy and associated non-CDI health expenditure due to treatment in 2020, but remains cost-effective even at a much
longevity. The incremental cost-effectiveness ratio of $4983 per higher cost, up to at least $6944.25,37
QALY is below the willingness-to-pay threshold of $100 000 per Based on our findings, we determine that sFMT with up to three
QALY, favoring sFMT. In both the 1-year and lifetime horizons, treatments remains cost-effective. Notably, only a small number of
sFMT is the preferred strategy. patients will require more than three treatments to achieve
Our model is validated by the close approximation of our 10 000 cure.10,13 We surmise that even if more than three FMTs are re-
iteration Monte Carlo simulation to recently described colectomy quired for a small proportion of patients, a reduced colectomy rate
and mortality outcomes. Figures 2 and 4 demonstrate that sFMT would offset the cost of additional FMTs and favor an sFMT strat-
is preferred well beyond the willingness-to-pay threshold of egy. We find that if an sFMT strategy is associated with decreased
$100 000, consistent with higher expected net benefit for sFMT, hospital length of stay in FCDI, sFMT for FCDI would be even
and indicate that there is little decision uncertainty regarding the more cost-effective than our model indicates. The value of infor-
preferred strategy. The cost-effectiveness of sFMT is likely due mation analysis demonstrates that the EVPI is quite low at $1.89

6 Journal of Gastroenterology and Hepatology •• (2021) ••–••


© 2021 Journal of Gastroenterology and Hepatology Foundation and John Wiley & Sons Australia, Ltd
S Gupta et al. Cost-effectiveness of fecal transplant

Figure 5 Scatterplot of results from 10 000 iterations of the probabilistic sensitivity analysis. Probabilistic sensitivity analysis results account for pa-
rameter uncertainties and show that the distribution of costs versus quality-adjusted life years (QALYs) overall favor a sequential fecal microbiota trans-
plantation strategy in the treatment of fulminant Clostridiodes difficile infection. , sequential fecal microbiota transplantation; , standard treatment.
[Color figure can be viewed at wileyonlinelibrary.com]

per person with an expected burden of removing decision uncer- model.11–13,15 Our analysis is also limited by lack of data to in-
tainty at $942 978. Given the lack of decision uncertainty that form risk of CDI recurrence after a treatment course, which has
we find, research at this cost or more may not impact treatment implications for long-term consequences of FMT. This is due to
strategy delineation. uncertainty regarding the risk of recurrent CDI after sFMT for pa-
There are multiple limitations to our study. First, and most nota- tients with FCDI. While current data suggest that FMT for recur-
bly, our parameter estimates rely heavily on one single-center pro- rent CDI is associated with lower rates of posttreatment
tocol with small sample size, using only data from patients with recurrence than medical therapy, this needs to be studied in pa-
FCDI and excluding patients with severe CDI.10 Our parameter ef- tients who undergo FMT for FCDI.12,41–45
fect sizes may be skewed by bias and may not be generalizable to The only risk of FMT we account for is possible perforation via
nonexpert centers; nevertheless, this prospective study was felt to colonoscopy. Research to ensure long-term safety of FMT is still
provide a robust basis for our model. This is confirmed by the val- being assessed, and so additional possible risks of FMT are not
idation of our outcomes in Monte Carlo simulation with recent, yet accounted for.46 It appears that the risk of adverse events asso-
large cohort data. The PSA helps account for parameter uncer- ciated with FMT is low, with one case report abstract of a bowel
tainty, which is also included in the value of information analysis. perforation associated with FMT.47,48 The limited details pub-
Because our findings suggest cost-effectiveness of sFMT despite lished regarding this case do not specify the method of FMT,
the large parameter uncertainty and even if the first FMT cure rates though the authors mention a protocol that appears to utilize
are low, sFMT may be beneficial in a community without an ex- FMT via nasogastric tube only.48–50 Thus, it is unclear if the re-
pert FMT provider. Furthermore, protocol heterogeneity and lack ported bowel perforation was indeed a consequence of FMT itself,
of specificity regarding outcomes per each sequential FMT in and we felt that perforation due to diagnostic colonoscopy better
FCDI alone in other studies precluded use of these data in our estimated this risk. Still, perforation during FMT is a risk that

Journal of Gastroenterology and Hepatology •• (2021) ••–•• 7


© 2021 Journal of Gastroenterology and Hepatology Foundation and John Wiley & Sons Australia, Ltd
Cost-effectiveness of fecal transplant S Gupta et al.

requires careful consideration in FCDI and may be more likely in Society for Healthcare Epidemiology of America (SHEA). Clin. Infect.
the absence of sufficient FMT expertise. Dis. 2018; 66: e1–48.
There are uncertainties regarding cost data. Because 5 Surawicz CM, Brandt LJ, Binion DG et al. Guidelines for diagnosis,
treatment, and prevention of Clostridium difficile infections. Am. J.
OpenBiome is currently the only commercially available product
Gastroenterol. 2013; 108: 478–98.
and further sources are not available for comparison, variations re-
6 Zhang S, Palazuelos-Munoz S, Balsells EM, Nair H, Chit A, Kyaw
garding stool preparation cost are difficult to assess. Additionally, MH. Cost of hospital management of Clostridium difficile infection in
the cost continues to increase, as safety considerations require rig- United States—a meta-analysis and modelling study. BMC Infect. Dis.
orous screening of stool.37 As OpenBiome can continue providing 2016; 16: 447.
FMT material for emergency cases such as FCDI despite the coro- 7 Luo Y, Lucas AL, Grinspan AM. Fecal transplants by colonoscopy and
navirus disease 2019 pandemic at listed prices, we do not antici- capsules are cost-effective strategies for treating recurrent
pate immediate price changes that would substantially impact Clostridioides difficile infection. Dig. Dis. Sci. 2019: 1–9.
our findings, especially because we find that sFMT is 8 Rajasingham R, Enns EA, Khoruts A, Vaughn BP. Cost-effectiveness
cost-effective up to almost $7000 per treatment. We did not in- of treatment regimens for Clostridioides difficile infection: an
evaluation of the 2018 Infectious Diseases Society of America guide-
clude the cost of rectal vancomycin, recommended for patients
lines. Clin. Infect. Dis. 2019.
with ileus with variable use.4 Given the low cost of medications
9 Varier RU, Biltaji E, Smith KJ et al. Cost-effectiveness analysis of fecal
compared with stool, colonoscopy, hospitalization, and colectomy, microbiota transplantation for recurrent infection. Infect. Control Hosp.
cost of possible rectal vancomycin would not likely change any Epidemiol. 2015; 36: 438–44.
outcome estimates. Lastly, our assumptions regarding cost of inpa- 10 Fischer M, Sipe B, Cheng Y-W et al. Fecal microbiota transplant in
tient colonoscopy may underestimate the real cost of the procedure severe and severe-complicated Clostridium difficile: a promising
in the inpatient setting compared with the outpatient setting, which treatment approach. Gut Microbes 2017; 8: 289–302.
is otherwise unknown. 11 Cammarota G, Ianiro G, Magalini S, Gasbarrini A, Gui D. Decrease in
Our simulation model is grounded in a real-world protocol and surgery for Clostridium difficile infection after starting a program to
is the first to evaluate cost-effectiveness of FMT for FCDI and transplant fecal microbiota. Ann. Intern. Med. 2015; 163: 487–8.
consider adverse events of FMT via colonoscopy. The value of in- 12 Cammarota G, Masucci L, Ianiro G et al. Randomised clinical trial:
faecal microbiota transplantation by colonoscopy vs. vancomycin for
formation analysis indicates that despite our reliance on limited
the treatment of recurrent Clostridium difficile infection. Aliment.
data with wide uncertainty guiding parameter estimates of FMT Pharmacol. Ther. 2015; 41: 835–43.
for FCDI, there is little societal benefit to be gained by pursuing 13 Ianiro G, Masucci L, Quaranta G et al. Randomised clinical trial: faecal
a randomized clinical trial, which would be fraught with logistic, microbiota transplantation by colonoscopy plus vancomycin for the
cost, and ethical concerns. It is worth noting that recommendations treatment of severe refractory Clostridium difficile infection—single
for maximum medical therapy with oral vancomycin, intravenous versus multiple infusions. Aliment. Pharmacol. Ther. 2018; 48: 152–9.
metronidazole, and possibly rectal vancomycin for FCDI have not 14 Ianiro G, Murri R, Sciumè GD et al. Incidence of bloodstream
been guided by data from randomized controlled trials.4 Moreover, infections, length of hospital stay, and survival in patients with
given our findings, there may be greater risk from a healthcare sec- recurrent Clostridioides difficile infection treated with fecal microbiota
tor perspective with standard therapy and forgoing FMT. Despite transplantation or antibiotics: a prospective cohort study. Ann. Intern.
Med. 2019; 171: 695–702.
the current recommendation for FMT only in multiply recurrent
15 Tixier EN, Verheyen E, Ungaro RC, Grinspan AM. Faecal microbiota
CDI, we expect that more clinicians will turn to FMT in FCDI, transplant decreases mortality in severe and fulminant Clostridioides
particularly as more data support its safety and efficacy. Our find- difficile infection in critically ill patients. Aliment. Pharmacol. Ther.
ings support the growing body of evidence that FMT can be con- 2019; 50: 1094–9.
sidered for the treatment of patients with FCDI. 16 Cheng Y-W, Phelps E, Nemes S et al. Fecal microbiota transplant
decreases mortality in patients with refractory severe or fulminant
Clostridioides difficile infection. Clin. Gastroenterol. Hepatol. 2020.
Acknowledgment 17 Mora Pinzon MC, Buie R, Liou J-I et al. Outcomes of community and
healthcare-onset Clostridium difficile infections. Clin. Infect. Dis.
The authors would like to thank Professor Myriam Hunink for her 2019; 68: 1343–50.
support. 18 Zilberberg MD, Shorr AF, Micek ST, Doherty JA, Kollef MH.
Clostridium difficile-associated disease and mortality among the
elderly critically ill. Crit. Care Med. 2009; 37: 2583–9.
19 Rokas KE, Johnson JW, Beardsley JR, Ohl CA, Luther VP, Williamson
References JC. The addition of intravenous metronidazole to oral vancomycin is
associated with improved mortality in critically ill patients with
1 Lessa FC, Mu Y, Bamberg WM et al. Burden of Clostridium difficile Clostridium difficile infection. Clin. Infect. Dis. 2015; 61: 934–41.
infection in the United States. N. Engl. J. Med. 2015; 372: 825–34. 20 Wang Y, Schluger A, Li J, Gomez-Simmonds A, Salmasian H,
2 Adams SD, Mercer DW. Fulminant Clostridium difficile colitis. Curr. Freedberg DE. Does addition of intravenous metronidazole to oral
Opin. Crit. Care 2007; 13: 450–5. vancomycin improve outcomes in Clostridioides difficile infection?
3 Sailhamer EA, Carson K, Chang Y et al. Fulminant Clostridium Clin. Infect. Dis. 2019.
difficile colitis: patterns of care and predictors of mortality. Arch. Surg. 21 Panteris V, Haringsma J, Kuipers E. Colonoscopy perforation rate,
2009; 144: 433–9. mechanisms and outcome: from diagnostic to therapeutic colonoscopy.
4 McDonald LC, Gerding DN, Johnson S et al. Clinical practice Endoscopy 2009; 41: 941–51.
guidelines for Clostridium difficile infection in adults and children: 22 Peprah D, Chiu AS, Jean RA, Pei KY. Comparison of outcomes
2017 update by the Infectious Diseases Society of America (IDSA) and between total abdominal and partial colectomy for the management of

8 Journal of Gastroenterology and Hepatology •• (2021) ••–••


© 2021 Journal of Gastroenterology and Hepatology Foundation and John Wiley & Sons Australia, Ltd
S Gupta et al. Cost-effectiveness of fecal transplant

severe, complicated Clostridium difficile infection. J. Am. Coll. Surg. 37 Kassam Z, Dubois N, Ramakrishna B et al. Donor screening for fecal
2019; 228: 925–30. microbiota transplantation. N. Engl. J. Med. 2019; 381: 2070.
23 Dallas KB, Condren A, Divino CM. Life after colectomy for fulminant 38 Arias E, Xu J. United States Life Tables: 2017. 2019.
Clostridium difficile colitis: a 7-year follow up study. Am. J. Surg. 39 Stagnitti MN. National health care expenses per person in the US
2014; 207: 533–9. civilian noninstitutionalized population, 2014. 2016.
24 Micromedex. IBM Watson Health. Greenwood Village, Colorado, 40 Strong M, Oakley JE, Brennan A, Breeze P. Estimating the expected
USA: (electronic version). value of sample information using the probabilistic sensitivity analysis
25 OpenBiome. FMT preparation information. 2019. sample: a fast, nonparametric regression-based method. Med. Decis.
26 Procedure Price Lookup Comparison File. CMS 2019. Cited 24 Making 2015; 35: 570–83.
January 2021.Available from: https://www.cms.gov/Medicare/ 41 Perler BK, Chen B, Phelps E et al. Long-term efficacy and safety of
Medicare-Fee-for-Service-Payment/HospitalOutpatientPPS/Annual- fecal microbiota transplantation for treatment of recurrent
Policy-Files-Items/2019-Annual-Policy-Files Clostridioides difficile infection. J. Clin. Gastroenterol. 2020.
27 McGlone SM, Bailey RR, Zimmer SM et al. The economic burden of 42 Gough E, Shaikh H, Manges AR. Systematic review of intestinal
Clostridium difficile. Clin. Microbiol. Infect. 2012; 18: 282–9. microbiota transplantation (fecal bacteriotherapy) for recurrent
28 Tengs TO, Wallace A. One thousand health-related quality-of-life Clostridium difficile infection. Clin. Infect. Dis. 2011; 53: 994–1002.
estimates. Med. Care 2000: 583–637. 43 Brandt LJ, Aroniadis OC, Mellow M et al. Long-term follow-up of
29 Fryback DG, Dunham NC, Palta M et al. US norms for six generic colonoscopic fecal microbiota transplant for recurrent Clostridium
health-related quality-of-life indexes from the National Health difficile infection. Am. J. Gastroenterol. 2012; 107: 1079–87.
Measurement study. Med. Care 2007; 45: 1162. 44 Jorup-Rönström C, Håkanson A, Sandell S et al. Fecal transplant
30 Wilcox MH, Ahir H, Coia JE et al. Impact of recurrent Clostridium against relapsing Clostridium difficile-associated diarrhea in 32
difficile infection: hospitalization and patient quality of life. patients. Scand. J. Gastroenterol. 2012; 47: 548–52.
J. Antimicrob. Chemother. 2017; 72: 2647–56. 45 Mattila E, Uusitalo-Seppälä R, Wuorela M et al. Fecal transplantation,
31 Slobogean GP, O’Brien PJ, Brauer CA. Single-dose versus through colonoscopy, is effective therapy for recurrent Clostridium
multiple-dose antibiotic prophylaxis for the surgical treatment of difficile infection. Gastroenterology 2012; 142: 490–6.
closed fractures: a cost-effectiveness analysis. Acta Orthop. 2010; 81: 46 Kelly CR, Kim AM, Laine L, Wu GD. The AGA’s fecal microbiota
256–62. transplantation national registry: an important step toward
32 Hayes JL, Hansen P. Is laparoscopic colectomy for cancer cost- understanding risks and benefits of microbiota therapeutics.
effective relative to open colectomy? ANZ J. Surg. 2007; 77: 782–6. Gastroenterology 2017; 152: 681–4.
33 Brown C, Gibson PR, Hart A et al. Long-term outcomes of colectomy 47 Wang S, Xu M, Wang W et al. Systematic review: adverse events of
surgery among patients with ulcerative colitis. Springerplus 2015; 4: fecal microbiota transplantation. PLoS One 2016; 11.
573. 48 Obi O, Hampton D, Anderson T et al. Fecal microbiota transplant for
34 Punekar YS, Hawkins N. Cost-effectiveness of infliximab for the treatment of resistant C. difficile infection using a standardized
treatment of acute exacerbations of ulcerative colitis. Eur. J. Health protocol: a community hospital experience: 2172. Am. J.
Econ. 2010; 11: 67–76. Gastroenterol. 2014; 109: S629.
35 Hay JW, Smeeding J, Carroll NV et al. Good research practices for 49 Rubin TA, Gessert CE, Aas J, Bakken JS. Fecal microbiome
measuring drug costs in cost effectiveness analyses: issues and transplantation for recurrent Clostridium difficile infection: report on a
recommendations: the ISPOR Drug Cost Task Force report—part I. case series. Anaerobe 2013; 19: 22–6.
Value Health 2010; 13: 3–7. 50 Seekatz AM, Aas J, Gessert CE et al. Recovery of the gut microbiome
36 Batty M, Ippolito B. Mystery of the chargemaster: examining the role following fecal microbiota transplantation. MBio 2014; 5: e00893–14.
of hospital list prices in what patients actually pay. Health Aff. 2017;
36: 689–96.

Journal of Gastroenterology and Hepatology •• (2021) ••–•• 9


© 2021 Journal of Gastroenterology and Hepatology Foundation and John Wiley & Sons Australia, Ltd

You might also like