Table 4.

(a) Correlations between the presence of crescents and composite outcomes in patients who were treated with corticosteroids + cyclophosphamide/mycophenolate mofetil; (b) Correlations between 7.39% of glomeruli with crescents and composite outcomes in patients who were treated with corticosteroids + cyclophosphamide/mycophenolate mofetil.

Combined event (Cox regression)
	Univariate hazard ratio (95% CI)	p Value	Multivariable hazard ratio (95% CI)*	p Value
a. Crescents
No	1	0.01	1	0.02
Any	2.3 (1.216–4.344)		3.2 (1.246–8.025)
b. Crescents		0.03		0.03
No	1		1
<7.39%	1.6 (0.540–5.033)	0.4	2.0 (0.547–7.471)	0.3
≥7.39%	2.4 (1.275–4.705)	0.007	3.5 (1.364–9.030)	0.009

Note: CI: confidence interval; AI: NIH activity index; CI: NIH chronicity index; eGFR: estimated glomerular filtration rate.

^*Multivariable: multivariable with crescents + initial age (lg transformed), gender, eGFR, proteinuria (lg transformed), AI (0–4, 5–8, >8), and CI (0–2, 3–4, 5–8). There was an increase in the hazard ratio of CI from 1 (0–2) to 3.6 (when 5–8) in Table 4a. And there was also an increase in the hazard ratio of CI from 1 (0–2) to 4.0 (when 5–8) in Table 4b. The other covariates were not statistically associated with a decreased survival from a combined event. The combined event including death, end-stage renal disease, or 30% reduction from baseline of eGFR, to increase the rate of events and permit a valid multivariable analysis, as mentioned in the text.