Risk for Asymptomatic Household Transmission of Clostridioides Difficile Infection Associated With Recently Hospitalized Family Members

Aaron C. Miller; Alan T. Arakkal; Daniel K. Sewell; Alberto M. Segre; Sriram V. Pemmaraju; Philip M. Polgreen

Emerging Infectious Diseases. 2022;28(5):932-939.

Results

We identified a total of 142,125,247 enrollees with ≥2 family members enrolled in the same insurance plan for an entire month (Table 1), which resulted in just over 5.1 billion enrollment months that we could observe over the study period. Most (53.2%) households contained ≥4 persons in the same insurance plan. We identified a total of 224,818 CDI cases across 194,424 enrollees; 55.9% of cases occurred among female enrollees and 74.6% among enrollees >40 years of age. Of all CDI cases, 6,575 cases represented a possible C. difficile transmission that occurred within 60 days after hospitalization of a family member. After we removed enrollees who were exposed to a family member with diagnosed CDI or who were hospitalized themselves, 164,650 CDI cases remained, of which 3,871 represented a potential asymptomatic C. difficile transmission from a recently hospitalized family member.

We calculated CDI incidence rates of cases per 100,000 enrollment months and unadjusted IRRs by the various demographic and exposure groups (Table 2). Consistent with established CDI risk factors, we found CDI incidence was greater among female persons; persons >40 years of age, especially persons >65 years of age; persons with exposure to low-CDI-risk and high-CDI-risk antibiotics; and persons taking PPIs. Overall, the CDI incidence was ≈73% greater (IRR 1.73) among persons exposed to a recently hospitalized family member (incidence of 5.56 cases/100,000 enrollment months) than among persons who were not exposed to recently a hospitalized family member (incidence of 3.22 cases/100,000 enrollment months). At a bivariate level across nearly all enrollment characteristics, the CDI incidence rate was greater among enrollees in households with recently hospitalized family members (Table 3). CDI incidence increased monotonically across the various levels of within-household hospitalization from 3.22 cases/100,000 enrollment months for 0 days of within-household hospitalization to 8.73 cases/100,000 enrollment months for >30 total days of within-household hospitalization.

For stratified regression analyses, we divided enrollees into 357,348 enrollment-month strata based on different combinations of demographics, enrollment characteristics, and risk factors (Table 4). For each within-household hospitalization exposure group, we computed IRRs relative to the baseline group in which family members spent <1 day in the hospital during the previous 60 days. Compared with enrollees whose family members spent <1 day in the hospital, the IRR of CDI continuously increased across the exposure bins from 1.30 (95% CI 1.19–1.41) for persons with 1–3 days of within-family hospitalization up to 2.45 (95% CI 1.66–3.60) for those with >30 days of within-family hospitalization.

Known CDI risk factors also were associated with greater incidence. Antimicrobial drug exposure was associated with an increased CDI incidence rate; for low-CDI-risk antibiotics the IRR was 2.69 (95% CI 2.59–2.79), and for high-CDI-risk antibiotics IRR was 8.83 (95% CI 8.63–9.03). PPI usage was also associated with statistically significant CDI incidence, an IRR of 2.23 (95% CI 2.15–2.30). CDI incidence increased with age; relative to ages 0–17 years the IRR continuously increased from 1.71 (95% CI 1.65–1.78) for ages 18–40 years to 9.32 (95% CI 8.92–9.73) for ages >65 years. Female persons had a higher incidence compared with male persons (IRR 1.30, 95% CI 1.28–1.33). Households with an infant also had a higher CDI incidence than those without (IRR 1.51, 95% CI 1.44–1.58).

We performed a sensitivity analysis to determine whether our results were confounded by household-level susceptibility (Appendix Table 1). When we reversed the temporal order of hospital exposure, we found little evidence that our primary results can be explained by confounding due to CDI susceptibility among family members. The point estimates for our primary dose-response curve remained relatively unchanged and were considerably larger than the effect estimates associated with future hospital visits among family members.

As a second sensitivity analysis, we considered a 90-day exposure window for capturing recently hospitalized family members (Appendix Tables 2–4). In general, the results of the analysis using a 90-day exposure window were consistent with the 60-day window, and we noted a similar dose-response relationship between the total days of within-household hospitalization among recently hospitalized family members and risk for CDI. However, the magnitude of some of the point estimates was slightly attenuated using the 90-day window compared with the 60-day window. For example, the IRR for the 1–3 day within-family hospitalization category was 1.24 for the 90-day window, compared with 1.30 for the 60-day window. However, the CIs for both sets of analyses overlapped the point estimates of the other.

1 2 3 4

Next Section

Table 1. Baseline enrollment characteristics of families with multiple infected members using a 60-day exposure window in study of asymptomatic *Clostridioides difficile* transmission among household members, United States*
Characteristics	All enrollees. no. (%)	No. (%) episodes of index CDI diagnosis†	No. (%) cases of possible transmission after family member hospitalization
No. CDI cases	NA	224,818	6,575
No. enrollees	142,125,247 (100)	194,424 (100)	6,453 (100)
Age group at enrollment or CDI diagnosis, y
0–17	47,733,847 (33.6)	19,719 (8.8)	547 (8.3)
18–40	46,634,859 (32.8)	37,259 (16.6)	1,156 (17.6)
41–65	44,039,682 (31.0)	103,430 (46.0)	1,822 (27.7)
>65	3,716,859 (2.6)	64,410 (28.6)	3,050 (46.4)
Sex
M	70,485,475 (49.6)	99,133 (44.1)	2,798 (42.6)
F	71,639,772 (50.4)	125,685 (55.9)	3,777 (57.4)
Family size
2	36,598,138 (25.8)	134,644 (59.9)	4,166 (63.4)
3	29,857,746 (21.0)	36,236 (16.1)	905 (13.8)
4	40,705,784 (28.6)	34,559 (15.4)	839 (12.8)
5	21,536,725 (15.2)	13,517 (6.0)	409 (6.2)
>5	13,426,854 (9.4)	5,862 (2.6)	256 (3.9)

*CDI, Clostridioides difficile infection; NA, not applicable.
†Events occurring ≥60 days before another episode.

Table 2. Bivariate comparisons of unadjusted incidence rates and incidence rate ratios for infection incidence across various patient strata using a 60-day exposure window in study of asymptomatic *Clostridioides difficile* transmission among household members, United States*
Variable	Exposed to previously hospitalized family member ≤60 days			Not exposed to previously hospitalized family member ≤60 days			Unadjusted IRR
Variable	CDI cases	Total enrollee months	CDI incidence†	CDI cases	Total enrollee months	CDI incidence†	Unadjusted IRR
Overall	3,871	69,675,026	5.56	160,779	4,998,101,178	3.22	1.73
Age group, y
0–17	317	24,432,280	1.30	15,615	1,445,786,086	1.08	1.20
18–40	567	19,978,891	2.84	29,718	1,427,785,479	2.08	1.37
41–65	1,193	19,281,059	6.19	74,803	1,868,106,655	4.00	1.55
>65	1,794	5,982,798	29.99	40,643	256,422,958	15.85	1.89
Sex
M	1,698	37,945,564	4.47	67,378	2,488,714,427	2.71	1.65
F	2,173	31,729,463	6.85	93,401	2,509,386,752	3.72	1.84
Outpatient antimicrobial drug use within 60 days
None	2,419	63,230,032	3.83	100,792	4,575,861,567	2.20	1.74
Low-risk drugs	292	2,979,748	9.80	11,944	201,200,918	5.94	1.65
High-risk drugs	1,160	3,465,248	33.48	48,043	221,038,693	21.74	1.54
PPI use within 30 days
N	3,477	68,273,806	5.09	146,185	4,913,346,960	2.98	1.71
Y	394	1,401,221	28.12	14,594	84,754,218	17.22	1.63
Infant age <2 y in family
N	3,489	48,618,765	7.18	151,291	4,597,497,625	3.29	2.18
Y	382	21,056,262	1.81	9,488	400,603,553	2.37	0.76

*IRRs compare CDI incidence among persons exposed to a family member previously hospitalized for ≥1 d relative incidence for to those not exposed to a previously hospitalized family member. The overall incidence rate ratio among those exposed to a previously hospitalized family member relative to those unexposed was 1.73 and was >1 across all strata. CDI, Clostridioides difficile infection; IRR, incidence rate ratio; PPI, proton-pump inhibitor.
†Cases per 100,000 enrollee months.

Table 3. Number of cases and enrollee-months in each exposure bin for total days of household-hospitalization using a 60-day exposure window in study of asymptomatic *Clostridioides difficile* transmission among household members, United States*
No. days family members spent hospitalized	60-day exposure window
No. days family members spent hospitalized	No. CDI cases	Total enrollment months	Incidence†
0	160,267	4,980,648,694	3.22
1–3	2,336	52,798,719	4.42
4–10	1,519	27,457,461	5.53
11–20	315	4,338,929	7.26
21–30	107	1,317,610	8.12
>30	106	1,214,792	8.73

*CDI, Clostridioides difficile infection.
†Cases per 100,000 enrollment months.

Table 4. Results of regression analysis of incidence rate ratio for *Clostridioides difficile* infection using quasi-Poisson model and 60-day exposure window in study of asymptomatic *C. difficile* transmission among household members, United States*
Variable	IRR (95% CI)
No. days member was hospitalized within 60 d
0	Referent
1–3	1.30 (1.19–1.41)
4–10	1.46 (1.32–1.62)
11–20	1.79 (1.43–2.23)
21–30	2.17 (1.48–3.18)
>30	2.45 (1.66–3.60)
Age group, y
0–17	Referent
18–40	1.71 (1.65–1.78)
41–65	2.97 (2.86–3.08)
>65	9.32 (8.92–9.73)
Sex
M	Referent
F	1.30 (1.28–1.33)
Outpatient antimicrobial drug use within 60 d
None	Referent
Low-risk drugs	2.69 (2.59–2.79)
High-risk drugs	8.83 (8.63–9.03)
PPI use within 30 d	2.23 (2.15–2.30)
Infant <2 y in family	1.51 (1.44–1.58)

*Models were adjusted for year, month, and family size. Regression models included an offset for number of enrollment months. Because family hospitalization exposure group was followed for 60 days to identify secondary Clostridioides difficile infection, the length of their enrollment period is 60 days. For the unexposed group, the length of enrollment was the length of a given month. IRR, incident rate ratio; PPI, proton-pump inhibitor.