Clinical course, treatment and outcome of Pneumocystis pneumonia in immunocompromised adults: a retrospective analysis over 17 years

Table 2 Regression model results for in-hospital mortality and overall survival in 240 patients with PcP underlining the role of LDH

Patient characteristic (potential predictors)^a	Logistic regression models for in-hospital mortality						Cox regression models for overall survival
	Univariate models			Multivariable model^b			Univariate models			Multivariable model^c
	Odds ratio	95% CI	p value	Odds ratio	95% CI	p value	Hazard ratio	95% CI	p value	Hazard ratio	95% CI	p value
LDH (50 U/L)	1.16	(1.10–1.24)	1.5 × 10⁻⁶	1.17	(1.09–1.27)	3.7 × 10⁻⁵	1.08	(1.06–1.11)	1.2 × 10⁻⁹	1.07	(1.04–1.10)	3.2 × 10^− 5
Age (5 years)	1.28	(1.15–1.44)	1.9 × 10^− 5	1.66	(0.82–3.87)	0.205	1.14	(1.05–1.24)	0.001	1.33	(0.76–2.32)	0.311
Sex (ref. male)	1.23	(0.65–2.31)	0.515				1.30	(0.81–2.09)	0.285
BMI (kg/m²)	1.14	(1.06–1.23)	1.5 × 10⁻⁴	0.66	(0.33–1.33)	0.229	1.11	(1.05–1.17)	1.7 × 10⁻⁴	0.65	(0.40–1.06)	0.082
Admission GFR (10 mL/min/1.73 m²)	0.90	(0.84–0.97)	0.003	0.80	(0.60–1.07)	0.121	0.96	(0.91–1.01)	0.084	1.01	(0.80–1.28)	0.914
Prednisone equivalent dose (50 mg/day)	1.52	(1.13–2.13)	0.010	1.32	(0.96–1.85)	0.083	1.08	(0.95–1.23)	0.231	1.04	(0.89–1.21)	0.626
TMP (10 kg/day)	0.70	(0.42–1.16)	0.166				0.87	(0.60–1.27)	0.469

Abbreviations: PcP Pneumocystis pneumonia, LDH lactate dehydrogenase, BMI body mass index, TMP trimethoprim-sulfamethoxazole, CI confidence interval
^aUnit for which effect was estimated or reference (ref.) in parenthesis
^bMultiple logistic regression including all predictors that met p ≤ 0.15 in one of the univariate analyses for both in-hospital mortality and overall survival; note that we also included quadratic terms for the continuous predictors (results not reported)
^cMultiple Cox regression including all predictors that met p ≤ 0.15 in one of the univariate analyses for both in-hospital mortality and overall survival; note that we also included quadratic terms for the continuous predictors (results not reported)

ISSN: 1364-8535