Role of Empiric Antifungal Therapy in Patients with Perforated Peptic Ulcers

Abstract

Background:

It is unclear if the addition of antifungal therapy for perforated peptic ulcers (PPU) leads to improved outcomes. We hypothesized that empiric antifungal therapy is associated with better clinical outcomes in critically ill patients with PPU.

Patients and Methods:

The 2001–2012 Medical Information Mart for Intensive Care (MIMIC-III) database was searched for patients with PPU and the included subjects were divided into two groups depending on receipt of antifungal therapy. Propensity score matching by surgical intervention, mechanical ventilation (MV), and vasopressor administration was then performed and clinically important outcomes were compared. Multiple logistic regression was performed to calculate the odds of a composite end point (defined as “alive, hospital-free, and infection-free at 30 days”).

Results:

A total of 89 patients with PPU were included, of whom 52 (58%) received empiric antifungal therapy. Propensity score matching resulted in 37 pairs. On logistic regression controlling for surgery, vasopressors, and MV, receipt of antifungal therapy was not associated with higher odds (odds ratio [OR], 1.5; 95% confidence interval [CI], 0.5–4.7; p = 0.4798) of the composite end point.

Conclusions:

In critically ill patients with perforated peptic ulcer, receipt of antifungal therapy, regardless of surgical intervention, was not associated with improved clinical outcomes. Selection bias is possible and therefore randomized controlled trials are required to confirm/refute causality.

Perforated peptic ulcer (PPU) is a potentially life-threatening complication of peptic ulcer disease (PUD) and a surgical emergency with substantial morbidity and mortality [1]. The incidence and prevalence of PUD and its complications have been declining mainly because of increased use of acid-suppression therapy and antimicrobial agents targeting Helicobacter pylori [2]. Perforated peptic ulcer has a lifetime prevalence of 5%, with 2%–14% of uncomplicated ulcers eventually perforating [1]. Despite a reduction in PUD complications, mortality rates in patients with PPU remain as high as 25%–30%, the majority of which are caused by sepsis [3].

Most patients undergoing operative management of PPU receive peri-operative antibiotic therapy given the contaminated nature of the procedure, however, utilization of antifungal therapy remains controversial. Although some studies associated the presence of fungal isolates in peritoneal fluid with worse clinical outcomes [4], others demonstrated that intra-peritoneal fungal growth is associated with older age and is not independently associated with outcomes [5]. Horn et al. [6] analyzed 7,000 patients with PPU undergoing surgery and found no differences in clinical outcomes between those treated pre-operatively with antifungal therapy and those who were not. In another multicenter study, empiric antifungal therapy failed to provide benefit in preventing the development of organ/space infection [7]. Although the evidence from these studies suggests that both pre-operative and post-operative antifungal therapy are not associated with better patient outcomes, there are no studies that included both operative and non-operative PPU in their analyses. We sought to explore the associations between antifungal therapy and clinical outcomes in critically ill patients with PPU regardless of operative intervention.

Patients and Methods

Data source

This study is a retrospective analysis of the Medical Information Mart for Intensive Care III (MIMIC-III) database, which is freely available and comprises 61,532 intensive care unit (ICU) admissions at Beth Israel Deaconess Medical Center from June 2001 through October 2012 [8,9]. The deidentified MIMIC-III data are compliant with the Health Insurance Portability and Accountability Act of 1996 (HIPAA) and all investigators with data access (M.C.) were approved by PhysioNet.

Baseline and outcome variables

We searched MIMIC-III for patients aged ≥18 years with PPU according to the International Classification of Diseases, Ninth Revision (ICD-9) codes for gastric (531.10, 531.11, 531.20, 531.21, 531.50, 531.51, 531.60, 531.61), duodenal (532.10, 532.11, 532.20, 532.21, 532.50, 532.51, 532.60, 532.61), gastrojejunal (534.10, 534.11, 534.20, 534.21, 534.50, 534.51, 534.60, 534.61), and unspecified PPU (533.10, 533.11, 533.20, 533.21, 533.50, 533.51, 533.60, 533.61). Baseline characteristics included age, gender, weight, cardiovascular comorbidities, lactic acid level on admission, vasopressor therapy on admission, mechanical ventilation on admission, operative management, and antifungal duration. Patients who met our inclusion criteria were included regardless if surgical intervention or non-operative management was utilized, and were then divided into two groups depending on receipt of antifungal therapy. We then compared clinically important outcomes between the two matched groups: hospital length of stay (LOS), 30-day hospital-free days, ICU LOS, ventilator days, 30-day ventilator-free days, intra-abdominal abscess, and in-hospital mortality. In addition, we defined a composite outcome “alive, hospital-free, and infection-free at 30 days.”

Statistical analyses

The Wilcoxon rank sum test was used to compare the non-parametric continuous outcomes and χ² test compared the categorical outcomes. Logistic regression was used to estimate propensity scores for our two groups and the nearest neighbor method (MatchIt package in R) was utilized to perform propensity score matching, while controlling for surgical intervention, mechanical ventilation (MV), and vasopressor administration. Multiple logistic regression was performed to calculate the odds and 95% confidence interval of the composite end point. A p value of <.05 was considered statistically significant.

Results

Demographics

A total of 89 patients with PPU were retrieved from the MIMIC database. The median (interquartile range [IQR]) age of the cohort was 64 [53,78] years and there was a relatively equal distribution of genders (Table 1). Most patients (61%) had the duodenal subtype of PPU. Of the total of 89 patients, 50 (56%) had undergone surgery and 18 (20%) were mechanically ventilated on admission. Fifty-two (58%) patients had received empiric antifungal therapy with either fluconazole, caspofungin, micafungin, or amphotericin B. There were no differences in demographics between patients who received empiric antifungal therapy and those who did not. The mean duration of antifungal therapy was seven days [4,12].

Table 1.

Demographics of Patients with Perforated Peptic Ulcer by Antifungal Therapy

		Antifungal therapy
	Total (n = 89)	No antifungal (n = 37)	Antifungal (n = 52)	p
Age (y)	64 [53–78]	62 [52–77]	65 [55–78]	0.54
Gender Female (%)	44 (49)	20 (54)	24 (46)	0.60
Weight (kg)	80.2 [70.1–97.4]	75.0 [69.0–90.0]	84.0 [71.6–100.0]	0.27
Comorbidities (%) Hypertension Diabetes mellitus Coronary artery disease	32 (36) 9 (10) 9 (10)	11 (30) 3 (8) 2 (5)	21 (40) 6 (12) 7 (14)	0.42 0.73 0.30
Perforated ulcer type (%) Gastric Duodenal Gastrojejunal (marginal) Unspecified	28 (31) 54 (61) 6 (7) 1 (1)	9 (24) 25 (68) 3 (8) 0	19 (36) 29 (56) 3 (6) 1 (2)
Lactic acid level on admission (mg/dL)	1.7 [1.4–3.4]	1.5 [1.0–3.4]	1.7 [1.4–3.3]	0.51
Vasopressors on admission (%)	10 (11)	2 (5)	8 (15)	0.18
Mechanically ventilated on admission (%)	18 (20)	5 (14)	13 (25)	0.28
Surgery (%)	50 (56)	19 (51)	31 (60)	0.58
Antifungal duration (days)	7 [4–12]		7 [4–12]

Data are presented as median [IQR] for non-parametric continuous variables and frequency (%) for categorical variables.

Outcomes

Propensity score matching between the antifungal and no antifungal groups resulted in 37 patients in each group (Table 2). Patients who received antifungal therapy had longer hospital and ICU LOS. There were no differences in the occurrence of intra-abdominal abscesses and mortality between the two groups. On logistic regression controlling for surgery, vasopressors, and mechanical ventilation, receipt of antifungal therapy was not associated with higher odds (OR, 1.5; 95% confidence interval [CI], 0.5–4.7; p = 0.4798) of the composite end point “alive, hospital-free, and infection-free at 30 days.”

Table 2.

Outcomes of Patients with Perforated Peptic Ulcer by Antifungal Therapy

	Propensity score natched
	No antifungal (n = 37)	Antifungal (n = 37)	p
Hospital LOS (days)	9 [6–13]	16 [10–24]	0.001
30-day hospital-free days	20 [0–23]	9 [0–17]	0.007
ICU LOS (days)	3 [1–5]	6 [2–10]	0.03
Ventilator days	0 [0–0]	0 [0–1]	0.001
30-day ventilator-free days	30 [30–30]	30 [29–30]	0.001
Intra-abdominal abscess (%)	4 (11)	4 (11)	1
In-hospital mortality (%)	10 (27)	7 (19)	0.58
Alive, hospital-free, and infection-free at 30 days (%)	23 (62)	22 (59)	1

LOS = length of stay; ICU = intensive care unit.

Data are presented as median [IQR] for non-parametric continuous variables and frequency (%) for categorical variables.

Discussion

In this retrospective analysis of a single-center ICU database, we report that empiric antifungal therapy is not associated with better clinical outcomes in both surgical and non-surgical PPU patients. The clinical benefits of antifungal agents to treat PPU have long been debated and randomized controlled trials (RCT) on this topic are few. One double-blind, multicenter RCT in Norway studied the effect of a single dose of fluconazole at the time of surgery in patients with intra-abdominal perforations [10]. The authors found that positive intra-operative fungal cultures were associated with higher morbidity and mortality, however, administration of intra-operative single-dose fluconazole did not improve survival. Although the World Society of Emergency Surgery and the Surgical Infection Society (SIS) agree on the benefits of antifungal therapy in high-risk surgical patients with intra-abdominal infections, their studies were not specific to patients with PPU [11,12]. In an Update Symposium at its 37th Annual Meeting, the SIS cautioned against the use of antifungal agents in patients with PPU because of lack of current evidence [13].

Our findings support previous studies regarding the limited use of antifungal therapy in patients with PPU. In a secondary analysis of a multicenter, case-control study, Barmparas et al. [7] found that use of empiric antifungal agents was not associated with decreased odds for organ/space infections in patients undergoing surgery for PPU. An interesting finding in this study was the wide variation of empiric antifungal use among the participating institutions, which highlights the necessity for universal recommendations on the use of these agents. Another retrospective study in Taiwan analyzed outcomes of patients with PPU-associated peritonitis with positive intra-abdominal fungal cultures. Comparing propensity score matched patients who received either fluconazole or an echinocandin for at least three days with those who did not, the study revealed no difference in 30-day all-cause mortality, the need for re-operation, and abscess formation [14]. Although similar, our study extended the findings to critically ill patients with PPU managed non-operatively and examined a composite end point comprising clinically relevant, patient-centered outcomes.

In addition to the significance in decreasing the risk of drug-induced adverse events, physicians should be mindful of antimicrobial resistance when prescribing these agents because of the risk of selection for more resistant species. In its 2019 report, the U.S. Centers for Disease Control and Prevention (CDC) cautioned the global health community on the emergence of multi-drug–resistant (MDR) fungal strains of Candida auris and azole-resistant Aspergillus fumigatus. The report demonstrated that the number of reported Candida auris cases increased by 318% in 2018 compared with 2015–2017 [15]. Other studies across multiple hospitals showed that 58%–76% of Candida auris infections resulted in invasive blood stream infections [16 –18]. Although research on antibiotic resistance has been pervasive and well-documented in recent years, little attention has been given to antifungal resistance. The Mycoses Study Group Education and Research Consortium recommends that antifungal stewardship goals should be implemented into hospital strategic policies with dedicated engagement of prescribing physicians [19]. In the modern era, prescribing antifungal agents for PPU is largely physician-dependent, thus further RCTs and cost-benefit analyses are needed to identify specific subpopulations and comorbidities of patients with PPU who may benefit from empiric therapy.

There are several limitations to our study that must be acknowledged. First, the study design is a post hoc analysis of a database and therefore subject to all the attendant biases. We attempted to minimize the impact of confounding variables by conducting propensity score matching, although selection bias is possible. In our study, patients who received antifungal therapy had significantly longer hospital and ICU LOS. This likely represents a selection bias where patients in the antifungal group had a more complicated hospital stay because of more severe disease at presentation compared with the no antifungal group. As in all retrospective studies, we can only describe associations and cannot prove or imply causality. Second, the small sample size at a single center increases the risk of type II error and limits reproducibility. Despite these limitations, our study does represent a meaningful contribution to the corpus of existing literature on this topic. Our study confirms and validates the findings of previous studies regarding the limited benefit of empiric antifungal therapy in these patients and extends to the population of critically ill patients with PPU managed non-operatively. Additionally, the incident rates described in our study can be used to perform informed power calculations for future randomized controlled trials. For example, using the incident rate of the composite endpoint (59%) in the antifungal group, a non-inferiority trial comparing antifungal to no-antifungal treatment with a 10% non-inferiority margin would require 415 subjects per arm.

Conclusions

The use of empiric antifungal therapy is not associated with a significant clinical advantage in critically ill patients with PPU managed operatively and non-operatively. More studies are required to establish consensus on the appropriate indications for antifungal agents in specific patient populations.

Footnotes

Funding Information

No funding was received.

Author Disclosure Statement

No competing financial interests exist.

References

Chung

, Shelat

. Perforated peptic ulcer: An update. World J Gastrointest Surg, 2017; 9:1–12.

Lau

, Sung

, Hill

, et al. Systematic review of the epidemiology of complicated peptic ulcer disease: Incidence, recurrence, risk factors and mortality. Digestion, 2011; 84:102–113.

Møller

, Adamsen

, Thomsen

, et al. Multicentre trial of a perioperative protocol to reduce mortality in patients with peptic ulcer perforation. Br J Surg, 2011; 98:802–810.

Shan

, Hsu

, Hsieh

, et al. Significance of intraoperative peritoneal culture of fungus in perforated peptic ulcer. Br J Surg, 2003; 90:1215–1219.

Kwan

, Lim

, Ng

, et al. Fungal Isolates in peritoneal fluid culture do not impact peri-operative outcomes of peptic ulcer perforation. Surg Infect, 2019; 20:619–624.

Horn

, Coleoglou Centeno

, Rasane

, et al. Pre-operative anti-fungal therapy does not improve outcomes in perforated peptic ulcers. Surg Infect, 2018; 19:587–592.

Barmparas

, Alhaj Saleh

, Huang

, et al. Empiric antifungals do not decrease the risk for organ space infection in patients with perforated peptic ulcer. Trauma Surg Acute Care Open. 2021; 6:e000662.

Johnson

AEW

, Pollard

, Shen

, et al. MIMIC-III, a freely accessible critical care database. Sci Data, 2016; 3:160035.

Goldberger

, Amaral

, Glass

, et al. PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals. Circulation, 2000; 101:e215–e220.

10.

Sandven

, Qvist

, Skovlund

, et al. Significance of Candida recovered from intraoperative specimens in patients with intra-abdominal perforations. Crit Care Med, 2002; 30:541–547.

11.

Mazuski

, Tessier

, May

, et al. The Surgical Infection Society Revised Guidelines on the Management of Intra-Abdominal Infection. Surg Infect, 2017; 18:1–76.

12.

Sartelli

, Catena

, Abu-Zidan

, et al. Management of intra-abdominal infections: Recommendations by the WSES 2016 consensus conference. World J Emerg Surg, 2017; 12:22.

13.

Huston

, Kreiner

, Ho

, et al. Role of Empiric anti-fungal therapy in the treatment of perforated peptic ulcer disease: Review of the evidence and future directions. Surg Infect, 2019; 20:593–600.

14.

, Lee

, Liu

. Antifungal therapy did not improve outcomes including 30-day all-cause mortality in patients suffering community-acquired perforated peptic ulcer-associated peritonitis with Candida species isolated from their peritoneal fluid. J Microbiol Immunol Infect, 2017; 50:370–376.

15.

CDC. Antibiotic Resistance Threats in the United States, 2019. Atlanta, GA: U.S. Department of Health and Human Services, CDC; 2019.

16.

Chowdhary

, Anil Kumar

, Sharma

, et al. Multidrug-resistant endemic clonal strain of Candida auris in India. Eur J Clin Microbiol Infect Dis, 2014; 33:919–926.

17.

Adams

, Quinn

, Tsay

, et al. Candida auris in healthcare facilities, New York, USA, 2013–2017. Emerg Infect Dis J, 2018; 24:1816.

18.

Sayeed

, Farooqi

, Jabeen

, et al. Clinical spectrum and factors impacting outcome of Candida auris: A single center study from Pakistan. BMC Infect Dis, 2019; 19:384.

19.

Johnson

, Lewis

, Dodds Ashley

, et al. Core recommendations for antifungal stewardship: A statement of the Mycoses Study Group Education and Research Consortium. J Infect Dis, 2020; 222(Suppl 3):S175–S198.