Skip to main content
. 2017 Oct;23(10):1671–1679. doi: 10.3201/eid2310.170797

Table 4. Cox proportional hazard regression analysis of potential independent variables associated with time to recurrence of Clostridium difficile infections, Hong Kong, China, 2006–2014*.

Variable Univariate analysis
Multivariate analysis
β Hazard ratio (95% CI) p value β Adjusted hazard ratio (95% CI) p value
Age, y
<44 NA 1.0 NA NA 1.0 NA
45−64 0.14 1.15 (0.86–1.55) 0.34 0.01 1.00 (0.74–1.36) 0.99
65−84 0.29 1.33 (0.01–1.75) 0.04 0.04 1.04 (0.78–1.38) 0.81
>85 0.41 1.50 (1.14–1.98) <0.01 0.16 1.17 (0.87–1.56) 0.29
Male sex −0.02 0.99 (0.88–1.10) 0.79 NA NA NA
Resident of home for elderly persons −0.02 0.98 (0.87–1.11) 0.99 NA NA NA
Severe disease† 0.32 1.38 (1.22–1.55) <0.01 0.35 1.41 (1.26–1.59) <0.01
Antimicrobial drug use‡
High-risk drug 0.49 1.55 (1.36–1.77) <0.01 0.32 1.37 (1.20–1.57) <0.01
Medium-risk drug 0.41 1.51 (1.33–1.72) 0.01 −0.01 0.99 (0.86–1.16) 0.96
Low-risk drug 0.11 1.12 (0.69–1.83) 0.66 NA NA NA
Diagnostic test
Bacterial culture NA 1.0 NA NA 1.0 NA
Toxin detection 0.69 1.89 (1.62–2.21) 0.01 0.57 1.79 (1.53–2.11) <0.01
NAAT 0.27 1.31 (1.12–1.52) 0.01 0.23 1.26 (1.08–1.47) <0.01
Use of proton-pump inhibitor 0.08 1.09 (0.97–1.22) 0.16 NA NA NA
Use of histamine-2 receptor antagonist 0.18 1.19 (1.07–1.34) 0.01 0.09 1.09 (0.97–1.22) 0.15
Healthcare-associated disease 0.49 1.65 (1.15–2.35) 0.01 0.42 1.52 (1.06–2.20) 0.02
Concurrent condition
Myocardial infarction 0.09 1.10 (0.92–1.32) 0.32 NA NA NA
Cerebrovascular disease 0.49 1.63 (1.46–1.82) <0.01 −0.15 0.86 (0.74–0.99) 0.04
Chronic lung disease −0.05 0.95 (0.82–1.12) 0.56 NA NA NA
Diabetes mellitus 0.08 1.08 (0.94–1.24) 0.30 NA NA NA
Renal disease 0.05 1.05 (0.92–1.20) 0.47 NA NA NA
Nonmetastatic tumor −0.24 0.79 (0.69–0.90) 0.01 −0.15 0.86 (0.74–0.99) 0.04
AIDS −0.58 0.56 (0.18–1.74) 0.32 NA NA NA
Inflammatory bowel disease 0.14 1.15 (0.62–2.14) 0.66 NA NA NA

*NA, not applicable; NAAT, nucleic acid amplification test.
†Severe disease was diagnosed according to the according to criteria of Cohen et al (13). 
‡Antimicrobial drug use 8 weeks before diagnosis was stratified into high risk (floroquinolones, cephalosporins, and clindamycin); medium risk (penicillins, macrolides, and sulfonamides); and low risk (tetracyclines).