Application of survival tree analysis for exploration of potential interactions between predictors of incident chronic kidney disease: a 15-year follow-up study

Table 7 Cox analyses of patterns identified using survival tree in the female population, Tehran Lipid and Glucose Study (1999–2014)

Nodes	Number of cases/number of events	Pattern description	HRs (95% CI)	p-value
12	750/43	Age ≤ 45 year, eGFR > 83.5 ml/min/1.73 m²	Reference	–
4	419/247	Age ≤ 45 year, eGFR ≤ 65 ml/min/1.73 m²	15.18 (10.97–20.99)	< 0.001
6	157/41	Age ≤ 29 year, 65 ≤ eGFR ≤ 71 ml/min/1.73 m²	5.07 (3.30–7.77)	< 0.001
7	549/220	29 ≤ Age ≤ 45 year, 65 ≤ eGFR ≤ 71 ml/min/1.73 m²	8.78 (6.32–12.15)	< 0.001
9	656/166	Age ≤ 45 year, 71 ≤ eGFR ≤ 76 ml/min/1.73 m²	4.93 (3.52–6.90)	< 0.001
11	799/124	Age ≤ 45 year, 75 ≤ eGFR ≤ 83.5 ml/min/1.73 m²	2.72 (1.92–3.86)	< 0.001
15	152/74	45 ≤ Age < 48 year, eGFR ≤ 69 ml/min/1.73 m²	12.06 (8.28–17.57)	< 0.001
16	580/423	Age > 48 year, eGFR ≤ 69 ml/min/1.73 m²	27.25 (19.88–37.34)	< 0.001
18	241/97	Age > 45 year, eGFR > 69 ml/min/1.73 m², SBP ≤ 133 mmHg	8.95 (6.25–12.82)	< 0.001
19	140/80	Age > 45 year, eGFR > 69 ml/min/1.73 m², SBP > 133 mmHg	16.41(11.32–23.80)	< 0.001

Column 1 shows the number of terminal nodes of survival tree of Fig. 5 for female population. Column 2 displays the sample size and number of events in each terminal node. Column 3 represents the patterns of each terminal node. Column 4 shows the HR for each node compared to reference node with high survival probability
CI Confidence intervals, HRs Hazard ratio, SBP Systolic blood pressure, eGFR Estimated glomerular filtration rate

ISSN: 1479-5876