Incidence of Clostridioides difficile Infections in Republic of Korea: A Prospective Study With Active Surveillance vs. National Data From Health Insurance Review & Assessment Service

Abstract

Background

Since the emergence of hypervirulent strains of Clostridioides difficile, the incidence of C. difficile infections (CDI) has increased significantly.

Methods

To assess the incidence of CDI in Korea, we conducted a prospective multicentre observational study from October 2020 to October 2021. Additionally, we calculated the incidence of CDI from mass data obtained from the Health Insurance Review and Assessment Service (HIRA) from 2008 to 2020.

Results

In the prospective study with active surveillance, 30,212 patients had diarrhoea and 907 patients were diagnosed with CDI over 1,288,571 patient-days and 193,264 admissions in 18 participating hospitals during 3 months of study period; the CDI per 10,000 patient-days was 7.04 and the CDI per 1,000 admission was 4.69. The incidence of CDI was higher in general hospitals than in tertiary hospitals: 6.38 per 10,000 patient-days (range: 3.25–12.05) and 4.18 per 1,000 admissions (range: 1.92–8.59) in 11 tertiary hospitals, vs. 9.45 per 10,000 patient-days (range: 5.68–13.90) and 6.73 per 1,000 admissions (range: 3.18–15.85) in seven general hospitals. With regard to HIRA data, the incidence of CDI in all hospitals has been increasing over the 13-year-period: from 0.3 to 1.8 per 10,000 patient-days, 0.3 to 1.6 per 1,000 admissions, and 6.9 to 56.9 per 100,000 population, respectively.

Conclusion

The incidence of CDI in Korea has been gradually increasing, and its recent value is as high as that in the United State and Europe. CDI is underestimated, particularly in general hospitals in Korea.

Graphical Abstract

INTRODUCTION

Since the emergence of hypervirulent strains of Clostridioides difficile, the incidence of C. difficile infections (CDI) has increased significantly, particularly in North America and Europe.1 In 2013, the Centers for Disease Control and Prevention (CDC) announced CDI as a major threat to the United States (U.S.). For the effective management of CDI in hospitals, guidelines recommend implementing surveillance of CDI rates and providing feedback to hospitals or wards, especially during endemic and outbreak situations.2

The incidence of CDI varies among countries.3 In the United State, according to a meta-analysis conducted from 2010 to 2019, the hospital-onset CDI incidence rate was 8.3 cases per 10,000 patient-days.4 In contrast, a meta-analysis showed that, compared to that of the US or Europe, Asian countries had a lower CDI incidence rate of 5.3 per 10,000 patient-days.5 In retrospective studies conducted in Korea, the incidence of CDI ranged from 1.7 per 1,000 admissions in 2004 to 2.7 per 1,000 admissions in 2008.6,7 However, in studies involving active surveillance, the incidence of CDI ranged from 2.7 to 7.16 per 10,000 patient days in 2011–20128 and 2009,9 respectively. These findings indicate that the incidence of CDI varies according to the institution and the surveillance program in place.

The risk factors contributing to the development of CDI include antibiotic use, advanced age, and hospitalization.10 Despite experiencing a decrease of −0.3% in expenditure from 2011 to 2020, Korea still ranks third among the 17 Organisation for Economic Co-operation and Development (OECD) countries in terms of expenditure on the Anatomical Therapeutic Chemical Classification System (ATC code J, which are anti-infectives for systemic use.11 Furthermore, by 2025, Korea is expected to enter a super-aged society, with a rapid increase in the older person population, accounting for 20.6% of the total population.12 Therefore, the incidence of CDI is expected to continue to increase.

To assess the current incidence of CDI in Korea, we conducted a prospective multicentre observational study using a similar active surveillance program in 18 hospitals across the Korean Peninsula. Additionally, we calculated the incidence of CDI from mass data obtained from the Health Insurance Review and Assessment Service (HIRA) from 2008 to 2020.

METHODS

Participating hospitals

Through October 2020 to October 2021, active surveillance for CDI was conducted for three months at tertiary and general hospitals in Korea. The number of participating hospitals in each region was determined based on the population size of the region (Fig. 1). The participating hospitals had infectious disease specialists serving as investigators along with a laboratory system for diagnosing CDI.

Fig. 1. Geographic information of the participating hospitals. The size of the circle represents the population of each administrative district in 2020 by Korean Statistical Information Service.

Active surveillance program for diarrhoea and diagnosis of CDI

During the study period, investigators monitored all admitted patients to select those who had newly developed unformed stools, three or more times per day, by using an electronic surveillance program for diarrhoea or reviewing electronic medical records (EMRs). Patients with unexplained diarrhoea were tested for CDI. If CDI tests were not performed in patients with diarrhoea, the researchers would leave a note to the physicians in charge to perform the CDI test.

Definition

To diagnose CDI, glutamate dehydrogenase (GDH), nucleic acid amplification test (NAAT), or enzyme immunoassay (EIA) toxin A&B assays were performed according to the hospital property. CDI was confirmed when an infectious disease physician suspected CDI based on clinical characteristics, and the EIA toxin A&B assay or NAAT was positive. These diagnostic criteria were determined with some modifications from the Society for Healthcare Epidemiology of America/Infectious Diseases Society of America (SHEA/IDSA) guidelines.13

The incidence of CDI was estimated based on the number of CDI cases per 10,000 patient-days or per 1,000 admissions. All CDI cases were included in the incidence rate regardless of recurrence. Patients diagnosed with CDI who developed diarrhoea at least 72 hours after admission or within 2 months of the last discharge were considered to have healthcare-associated CDI (HA-CDI). Patients with CDI who did not meet the criteria for HA-CDI were considered community-acquired CDI (CA-CDI).14 The researchers investigated the medical history through medical records and interviews.

Incidence of CDI from the national data of the HIRA

We obtained big data from the HIRA from 2008 to 2020 with permission (M20211021591). All admitted patients aged ≥ 20 years who were diagnosed with A047 as the main or sub-diagnosis were included. According to the Korean Standard Classification of Diseases, A047 includes enteritis caused by C. difficile and pseudomembranous colitis. Patients who were admitted for 0 days and those with admission costs for 0 days were excluded from the study. As for the incidence rate per population, data from Statistics Korea were used as the denominator because the National Health Insurance System of Korea covers almost the entire population.15

Statistical analysis

Continuous variables were analysed using the Mann–Whitney U test to determine differences. The incidence trend was analysed using a linear regression model. Statistical analyses were performed using SPSS version 27 (IBM Corp., Armonk, NY, USA).

Ethics statement

This study was conducted in accordance with the guidelines of the Declaration of Helsinki. The study protocols were reviewed and approved by the Institutional Review Board of Hanyang University Hospital and written informed consent was obtained from all the participants (approval no. HYUH 2020-06-046 for active surveillance; HYUH 2021-02-019 for big data). All methods were performed in accordance with relevant guidelines and regulations.

RESULTS

Prospective observational study with active surveillance

Characteristics of the participating hospitals

Initially, 12 tertiary hospitals and seven general hospitals participated in this study (Fig. 1). A tertiary hospital is a teaching hospital affiliated with a medical school that provides a higher level of medical services and expertise than general hospitals. During the study period, one tertiary hospital was temporarily converted into a coronavirus disease 2019 specialized hospital, which made the demographic characteristics of the admitted patients different from those of other hospitals. This hospital was therefore excluded. Finally, 11 tertiary hospitals and seven general hospitals, were included. The characteristics of the participating hospitals are listed in Table 1.

Table 1. Characteristics of participating hospitals.

Variables		A	B	C	E	F	G	H	I	J	L	M	N	O	P	Q	R	S	T
Hospital type		T	T	T	T	T	T	T	T	T	G	T	G	G	G	G	G	G	T
Screening method		E	M	E	M	E	E	M	M	E	E	E	E	E	E	M	E	E	E
Capacity of hospital		855	895	1,205	684	799	894	1,371	1,084	1,334	725	700	670	613	463	315	648	402	762
Laboratory test for CDI
	GDH	O		O			O		O	O	O			O	O			O
	PCR	O	O	O	O	O		O	O		O	O	O	O	O	O	O	O	O
	EIA toxin A&B	O	O	O	O	O	O	O	O	O	O	O	O	O	O	O	O	O	O
	Culture	O			O	O			O	O		O		O	O		O

T = tertiary hospital, G = general hospital, E = electronic surveillance program, M = manual review of electronic medical records, CDI = Clostridioides difficile infection, GDH = glutamate dehydrogenase, PCR = polymerase chain reaction, EIA = enzyme immunoassay, O = Tests performed at the hospital.

Of these, 13 hospitals screened patients with diarrhoea using an electronic surveillance program for diarrhoea, and 5 hospitals manually reviewed the EMR for the detection of patients with diarrhoea. Among 18 hospitals, nine (50%) hospitals performed GDH, 15 (83.3%) hospitals performed polymerase chain reaction as NAAT, 18 (100%) hospitals performed EIA toxin A&B assay, and nine (50%) hospitals performed C. difficile culture tests.

Incidence of CDI

Within 1,288,571 patient-days and 193,264 admissions, 30,212 patients had diarrhoea and 907 patients were diagnosed with CDI. Therefore, the incidence was 7.04 per 10,000 patient-days and 4.69 per 1,000 admissions.

In the tertiary hospitals, 24,592 patients complained of diarrhoea. Among them, 646 patients (2.6%) were confirmed as having CDI. Total CDI/10,000 patient-days was 6.38 (range: 3.25–12.05) and total CDI/1,000 admissions was 4.18 (range: 1.92–8.59). Table 2 shows the incidence of CDI at each tertiary hospital. Among patients with CDI, a total of 39 (6%) complained of diarrhoea less than 72 hours after the admission. Therefore, the HA-CDI incidence rate in tertiary hospitals was 5.99 per 10,000 patient-days and 3.93 per 1,000 admissions.

Table 2. CDI incidence rates in participating hospitals.

Variables	Tertiary hospitals											General hospitals
	A	B	C	E	F	G	H	I	J	M	T	L	N	O	P	Q	R	S
No. of diarrheal patients	7,267	866	353	2,726	651	4,135	2,434	897	758	3,510	995	239	1,337	1,344	935	309	1,164	292
No. of CDI patients	151	26	56	28	44	39	54	85	57	65	41	53	37	43	21	29	48	30
No. of healthcare-associated CDI patients	136	25	56	26	44	36	50	80	57	60	37	53	30	38	17	24	47	23
No. of community-associated CDI patients	15	1	0	2	0	3	4	5	0	5	4	0	7	5	4	5	1	7
Patient-days	131,824	80,113	121,422	66,131	75,565	67,555	94,363	119,642	139,856	53,920	62,122	55,152	47,851	44,294	36,979	20,866	40,252	30,664
Admission	17,586	13,559	17,227	10,701	12,441	10,753	11,404	14,077	26,625	10,199	9,896	4,419	7,290	7,255	6,613	1,830	6,046	5,343
CDI/Patient-days*104	11.45	3.25	4.61	4.23	5.82	5.77	5.72	7.10	4.08	12.05	6.60	9.61	7.73	9.71	5.68	13.90	11.92	9.78
HA-CDI/Patient-days*104	10.32	3.12	4.61	3.93	5.82	5.33	5.30	6.69	4.08	11.13	5.96	9.61	6.27	8.58	4.60	11.50	11.68	7.50
CDI/admission*103	8.59	1.92	3.25	2.62	3.54	3.63	4.74	6.04	2.14	6.37	4.14	11.99	5.08	5.93	3.18	15.85	7.94	5.62
HA-CDI/admission*103	7.73	1.84	3.25	2.43	3.54	3.35	4.38	5.68	2.14	5.88	3.74	11.99	4.12	5.24	2.57	13.11	7.77	4.30
No. of 1st episode of CDI	137	25	54	26	42	37	52	81	54	59	34	49	36	43	19	26	42	28

CDI = Clostridioides difficile infection, HA-CDI = healthcare-associated CDI.

In the general hospitals, 5,620 patients complained of diarrhoea. Among these, 261 patients (4.6%) were confirmed as having CDI. Total CDI/10,000 patient-days was 9.45 (range: 5.68–13.90) and CDI/1,000 admissions was 6.73 (range: 3.18–15.85). Table 2 shows the incidence of CDI at each hospital. Among patients with CDI, 29 (11.1%) complained of diarrhoea within 72 hours after the admission. Therefore, the HA-CDI incidence rate in general hospitals was 8.4 per 10,000 patient-days and 5.98 per 1,000 admissions. Therefore, as for the origin of CDI, most were HA-CDI (92.5%), and CA-CDI represented a low proportion (7.5%): 6.03% in tertiary hospitals and 11.1% in general hospitals (Table 2).

Comparing the CDI incidences between the hospitals with electronic surveillance programs for diarrhoea and those using manual screening of EMR, the incidence rates of CDI in the former hospitals tended to be higher than those in the latter hospitals, but without statistical significance. The median CDI incidence rate was 5.72 per 10,000 patient-days and 4.74 per 1,000 admissions by the manual screening method and 7.73 per 10,000 patient-days and 5.08 per 1,000 admissions by the electronic screening method (P = 0.349 and P = 0.73, respectively, by the Mann-Whitney U test).

Incidence of CDI from the national data of the HIRA from 2008 to 2020

Using national data from the HIRA, the annual incidences of CDI between 2008 and 2020 were measured and compared. The incidence of CDI was assessed per 10,000 patient days, 1,000 admissions, and 100,000 individuals. From 2008 to 2020, the incidence of CDI gradually increased significantly. Based on the CDI incidence per 10,000 patient-days, the incidence increased from 1 to 5.1 in tertiary hospitals (P < 0.001, adjusted R² = 0.955), and from 0.6 to 5 in general hospitals (P < 0.001, adjusted R² = 0.971) (Fig. 2). Based on the CDI incidence per 1,000 admissions, the incidence increased from 0.8 to 2.8 in tertiary hospitals (P < 0.001, adjusted R² = 0.949) and from 0.5 to 3.2 in general hospitals (P < 0.001, adjusted R² = 0.963). The CDI incidence per 100,000 individuals also increased significantly from 6.9 to 56.9 in all hospital types (P < 0.001, adjusted R² = 0.985). The gross incidence rate of CDI increased approximately 3-fold in tertiary hospitals, general hospitals, and all hospitals in Korea over a period of 10 years from 2011 (Fig. 2).

Fig. 2. CDI incidence according to hospital type in Korea using data from the Health Insurance Review & Assessment Service (HIRA) from 2008 to 2020: (A) per 10,000 patient-days, (B) per 1,000 admissions, (C) per 100,000 population years.

CDI = Clostridioides difficile infection.

When comparing the incidence rate of CDI between prospective active surveillance and retrospective national data analysis, the incidence rate was underestimated in the national data analysis, particularly in general hospitals.

DISCUSSION

In this study, a prospective multicentre observational study using an active surveillance program, and an analysis of the data from the HIRA, we demonstrated that the incidence of CDI in Korea was as high as that in the U.S. and other European countries. This is in contrast to previously published studies where the value was found to be lower than that in Western countries and similar to that in other Asian countries.3,16,17

Furthermore, the incidence of CDI in Korea has gradually increased over the past 13 years. In a previous prospective study conducted in 2009 at a single centre in Korea, the incidence rate of CDI was 7.16 cases per 10,000 patients-day,9 and the value in the same hospital increased to 11.45 cases per 10,000 patients-days in 2020–2021 in this study. There may be several reasons for the increase in CDI incidence: First, the older adult population has been increasing, and broad-spectrum antibiotics have been used more widely in Korea.11,12 Second, physicians recognised CDI considerably more than they did a decade ago, and tested for CDI more frequently. Third, the diagnostic method has changed; the EIA toxin A&B assay has been the most common diagnostic tool for CDI in Korea, and the NAAT or GDH tests have been introduced and covered by insurance since 2017 in Korea18; however, the CDI incidence might be overestimated because of the high detection rate of NAAT.19

According to the SHEA/IDSA guidelines, a suitable stool sample could be diagnosed as CDI by NAAT alone or by the stool toxin test as part of a multiple step algorithm.13 We applied these guidelines to diagnose CDI when NAAT was positive, or when GDH and toxins were positive. However, the GDH test was performed in only 50% of participating hospitals, making it difficult to recommend as a diagnostic standard. Therefore, only toxin test results without GDH were included in the diagnostic criteria, in this study. Confirming the diagnosis of CDI using a single test is challenging. The key to diagnosing CDI is to suspect it in patients experiencing unformed stools three or more times per day without a clear cause, ordering the necessary diagnostic tests for the diagnosis of CDI, and performing the tests appropriately under anaerobic conditions.

While we collected the CDI cases among hospitalised patients, 6.03% and 11.1% of patients with CDI in tertiary and general hospitals, respectively, were defined as CA-CDI. The proportion of CA-CDI in this study was lower than that reported in the U.S. or Europe17,20,21 and we considered that the proportion of CA-CDI in our study was underestimated. Since we enrolled patients with diarrhoea among hospitalised patients who were not in outpatient clinics or emergency rooms, patients with CA-CDI who were not hospitalised may have been missed in this study. Additionally, the participating hospitals were general or tertiary hospitals, and patients with mild CA-CDI may have visited primary outpatient clinics. Therefore, to determine the approximate incidence of CA-CDI in Korea, a prospective study including outpatient clinics and emergency units is necessary.

The incidence rate of CDI in hospitals is dependent on the appropriate detection of patients with diarrhoea as well as using appropriate tests for CDI. As expected, the incidence rates of active surveillance were higher than those of the HIRA, especially in general hospitals. To detect patients with diarrhoea without fail, a hospital in the U.S. reported the adoption of a policy allowing nurses to independently order stool samples for new patients with CDI symptoms.22 In this study, 64% of tertiary hospitals and 86% of general hospitals employed electronic surveillance programs to detect patients with diarrhoea, which successfully reduced the workload of the researchers. The programs for surveying diarrhoea were easily created in each hospital using an EMR system. The incidence of CDI tended to be higher in hospitals with electronic surveillance programs for diarrhoea; however, this difference was not statistically significant. Even with this screening system in place, screening all relevant patients is difficult, which might partly explain the difference in incidence rates among hospitals.

According to the Korea National Antimicrobial Use Analysis System data for 2021, antibiotic usage in general hospitals is higher than that in tertiary hospitals. Indeed, the incidence rate of CDI in the general hospitals, participating in the study, was higher than that in the tertiary hospitals. However, national data have shown that the CDI incidence rate in secondary hospitals is lower than that in tertiary hospitals. The lack of awareness of CDI among medical staff, poor detection of relevant patients with diarrhoea, and lack of available CDI tests might contribute to the low incidence rate in general hospitals.

Hospitalisation is an important risk factor for the acquisition of CDI. As the CDI incidence rate and average length of stay (ALS)23 in each hospital showed two- to three-fold differences among the participating hospitals, we analysed the association between the turnover rate of beds (TRB)24 or ALS and CDI incidence rates in the participating hospitals. The incidence rate of CDI was negatively correlated with TRB and positively correlated with ALS (data not shown). However, the incidence of CDI in hospitals is influenced by many factors, including antibiotic usage and hospital infection control (environmental cleaning and isolation of patients with CDI). Therefore, surveying multiple risk factors for CDI in hospitals is necessary to determine the relationship between the incidence rate of CDI and TRB or ALS.

Despite the overall increase in CDI worldwide, England and the United States have recently shown a decrease in the incidence of HA-CDI.25,26 Particularly in England, there has been an approximately 80% decrease in incidence since 2006. The most probable reason for this decline is the restriction of fluoroquinolone and cephalosporin use, which showed a high correlation with CDI occurrence. Additionally, C. difficile is a spore-forming bacterium that exhibits resistance to alcohol and can survive for long periods in the environment, making it a potential reservoir for persistent infections and reinfection.10 Therefore, infection control is important for decreasing the risk of C. difficile transmission in hospitals.27 In Korea, the predominant form of CDI is largely associated with HA-CDI. For the correct management of CDI, reinforced antibiotic stewardship and multidisciplinary infection control measures are necessary.

In conclusion, the incidence of CDI in Korea increased from 2008 to 2020, and its recent value was as high as that in the U.S. and Europe. CDI is underestimated, particularly in general hospitals in Korea.