Table 2. Univariate predictors of CDI in the biospecimen group and observational group, by Cox proportional hazards regression.

Predictor	Biospecimen group (N = 94)a		Observational Group (N = 1144)b
	Haz ratio (95% CI)	P	Haz ratio (95% CI)	P
Age (years)	0.98 (0.94–1.02)	0.311	1.01 (0.99–1.02)	0.279
Sex (female)	0.41 (0.12–1.14)	0.089	1.02 (0.72–1.42)	0.928
Underlying Disease (leukemia vs. other)	2.44 (0.92–7.35)	0.075	1.42 (1.01–2.02)	0.044
Conditioning Regimen (myeloablative vs. other)	3.18 (1.14–10.63)	0.026	1.99 (1.30–3.17)	0.001
T-cell depleted graft	1.96 (0.72–5.51)	0.185	1.59 (1.14–2.24)	0.007
Stem cell source (cord vs. other)	0.49 (0.11–1.62)	0.263	1.31 (0.74–2.16)	0.331
Prior antibiotics (14 days) f	1.31 (0.50–3.45)	0.580
Antibiotics c
Vancomycin (IV)	3.16 (0.85–11.91)	0.085	0.79 (0.53–1.18)	0.242
Metronidazole	0.43 (0.00–3.63)	0.518	0.73 (0.41–1.24)	0.252
Fluoroquinolones d	0.28 (0.03–1.44)	0.139	0.90 (0.60–1.33)	0.605
Beta-lactam e	1.28 (0.40–3.78)	0.666	0.67 (0.45–1.00)	0.048
tcdB positivity c	17.16 (6.40–51.97)	0.000

^aMultivariate analysis of the biospecimen group can be found in Table S2.

^bMultivariate analysis of the observational group can be found in Table S3.

^cAnalyzed as a time-varying predictor.

^dFluoroquinolones consist of ciprofloxacin and levofloxacin.

^eBeta-lactams include cephalosporins, beta-lactam/beta-lactamase combinations, and carbapenems.

^fPrior antibiotics refer to antibiotics given prior to allo-HSCT and prior to observation time, within 14 days, and were not analyzed as time-varying predictors.