Nomogram and Web Calculator Based on Lasso-Logistic Regression for Predicting Persistent Organ Failure in Acute Pancreatitis Patients

Figures & data

Figure 1 Flowchart of patient enrollment.

Table 1 Differences Between the Two Groups in AP Patients with or Without POF

Characteristics	Total (n=580)	Without POF (n=435)	With POF (n=145)	P-value
Sex:				0.194
Male	372 (64.1%)	286 (65.7%)	86 (59.3%)
Female	208 (35.9%)	149 (34.3%)	59 (40.7%)
Age (years)	48.0 [36.0;60.2]	46.0 [35.0;57.0]	56.0 [41.0;70.0]	<0.001*
BMI (kg/m^2)	25.9 [23.5;27.8]	25.9 [23.5;27.8]	26.0 [23.7;27.6]	0.785
Temperature (°C)	36.8 [36.5;37.0]	36.8 [36.5;37.0]	36.9 [36.6;37.4]	<0.001*
Heart Rate (beats/mins)	88.0 [78.0;104]	85.0 [77.0;98.0]	102 [85.0;124]	<0.001*
Respiratory Rate (beats/mins)	20.0 [20.0;21.0]	20.0 [20.0;21.0]	20.0 [20.0;23.0]	<0.001*
Systolic Blood Pressure (mmHg)	128 [118;142]	128 [118;140]	130 [117;144]	0.267
Diastolic Blood Pressure (mmHg)	78.0 [70.8;87.0]	78.0 [71.0;86.0]	78.0 [70.0;87.0]	0.969
Smoking:				0.355
No	339 (58.4%)	249 (57.2%)	90 (62.1%)
Yes	241 (41.6%)	186 (42.8%)	55 (37.9%)
Diabetes:				0.949
No	483 (83.3%)	363 (83.4%)	120 (82.8%)
Yes	97 (16.7%)	72 (16.6%)	25 (17.2%)
Hypertension:				0.012*
No	417 (71.9%)	325 (74.7%)	92 (63.4%)
Yes	163 (28.1%)	110 (25.3%)	53 (36.6%)
Hyperlipemia:				1.000
No	474 (81.7%)	356 (81.8%)	118 (81.4%)
Yes	106 (18.3%)	79 (18.2%)	27 (18.6%)
Fatty Liver:				0.086
No	435 (75.0%)	318 (73.1%)	117 (80.7%)
Yes	145 (25.0%)	117 (26.9%)	28 (19.3%)
ALT (IU/L)	40.0 [21.3;130]	41.6 [21.4;131]	35.0 [19.5;104]	0.307
AST (IU/L)	33.2 [21.7;110]	32.2 [20.5;111]	43.7 [27.0;92.0]	0.022*
Albumin (g/L)	36.6 [32.1;40.9]	37.5 [34.2;41.7]	32.2 [29.0;37.3]	<0.001*
Total Bilirubin (μmol/L)	24.1 [15.6;36.5]	23.8 [14.9;36.8]	25.3 [17.5;35.1]	0.376
Direct Bilirubin (μmol/L)	5.60 [3.27;12.9]	5.20 [3.00;12.2]	7.30 [4.30;13.2]	0.003*
Total Cholesterol (mmol/L)	4.11 [3.24;5.15]	4.23 [3.45;5.24]	3.77 [2.69;4.90]	<0.001*
Triglyceride (mmol/L)	1.33 [0.83;3.38]	1.34 [0.82;3.25]	1.30 [0.86;3.64]	0.863
Urea (mmol/L)	5.20 [4.00;6.90]	5.00 [4.00;6.30]	6.60 [4.70;10.0]	<0.001*
Serum Creatinine (μmol/L)	75.0 [61.4;89.1]	73.3 [60.3;85.9]	84.7 [66.3;109]	<0.001*
Amylase (IU/L)	189 [75.9;649]	153 [69.8;550]	329 [98.3;945]	0.001*
Lipase (IU/L)	233 [86.1;772]	213 [85.7;739]	386 [120;808]	0.044*
Serum Potassium (mmol/L)	3.93 [3.63;4.22]	3.93 [3.64;4.20]	3.92 [3.61;4.28]	0.947
Serum Sodium (mmol/L)	136 [134;139]	137 [134;139]	136 [133;139]	0.034*
Serum Chlorine (mmol/L)	103 [99.8;105]	103 [99.9;105]	103 [99.5;107]	0.058
Serum phosphorus (mmol/L)	0.97 [0.72;1.16]	1.01 [0.79;1.17]	0.82 [0.55;1.10]	<0.001*
Serum Magnesium (mmol/L)	0.82 [0.75;0.89]	0.82 [0.76;0.89]	0.80 [0.73;0.87]	0.027*
Serum Calcium (mmol/L)	2.15 [2.02;2.27]	2.18 [2.08;2.28]	2.03 [1.82;2.18]	<0.001*
WBC (×10^9/L)	11.2 [8.07;14.6]	10.7 [7.70;13.8]	13.2 [9.60;16.4]	<0.001*
Neutrophil count (×10^9/L)	82.2 [73.6;88.0]	80.6 [70.8;86.2]	87.2 [82.0;90.8]	<0.001*
Lymphocyte count (×10^9/L)	10.6 [6.30;17.0]	12.0 [7.70;19.4]	6.70 [4.70;10.9]	<0.001*
Monocyte count (×10^9/L)	5.70 [4.00;7.30]	5.90 [4.20;7.30]	5.00 [3.20;7.00]	0.001*
Eosinophils count (×10^9/L)	0.40 [0.10;1.00]	0.60 [0.10;1.10]	0.10 [0.00;0.60]	<0.001*
Basophils count (×10^9/L)	0.30 [0.10;0.50]	0.30 [0.20;0.50]	0.20 [0.10;0.40]	<0.001*
RBC (×10^12/L)	4.53 [4.07;5.01]	4.53 [4.10;4.97]	4.57 [4.03;5.15]	0.776
Hemoglobin (g/L)	143 (22.2)	142 (20.9)	144 (25.7)	0.437
Red cell distribution width (%)	13.3 [12.1;14.1]	13.2 [11.9;14.0]	13.7 [12.7;14.4]	0.001*
Platelet count (×10^9/L)	0.18 [0.14;0.22]	0.18 [0.15;0.22]	0.16 [0.13;0.21]	0.004*
Platelet distribution width (%)	16.6 [16.1;17.3]	16.6 [16.0;17.2]	16.9 [16.3;17.4]	0.004*
Mean platelet volume (fL)	8.60 [7.90;9.50]	8.50 [7.85;9.40]	8.80 [7.90;9.60]	0.049*
Prothrombin time (seconds)	12.6 [11.9;13.9]	12.6 [11.8;13.6]	13.2 [12.1;14.6]	<0.001*
International normalized ratio	1.17 [1.10;1.28]	1.16 [1.09;1.26]	1.22 [1.12;1.34]	<0.001*
D dimer (ng/mL)	490 [214;1254]	378 [177;866]	1269 [487;2813]	<0.001*
Activated partial thromboplastin time (seconds)	29.8 [28.1;32.3]	29.7 [28.1;32.2]	30.1 [27.9;33.4]	0.296

Notes: Value are presented as number (%), mean±standard deviation or IQR, 50 (25, 75). Normally distributed data were tested using independent samples t-tests, and nonparametric tests were used for nonnormal data. *P<0.05.

Abbreviations: IQR, interquartile range. BMI, body mass index; ALT, alanine aminotransferase, AST, aspartate aminotransferase; WBC, White Blood Cell Count; RBC, red blood cell.

Figure 2 Comparison of clinical outcomes based on whether there are complicated with POF. Length of stay. For non-normal distributed data, Mann–Whitney test was used. P<0.001. ****P<0.001. (B) ICU stay. (C) Bacterial Infection. (D) Death. (E) Walled-off necrosis. (F) Pancreatic fluid collections. (B and D) Fisher’s exact test. P<0.001. (C, E and F) Chi-square test. P <0.001.

Table 2 Multivariate Logistic Regression Results

Variable	β	CI	P
Age	0.0482	1.049 (1.033–1.066)	<0.001
Heart Rate	0.0368	1.037 (1.025–1.050)	<0.001
Serum Creatinine	0.0076	1.007 (1.002–1.013)	0.008
Lymphocytes	−0.0480	0.953 (0.919–0.985)	0.007
Serum Calcium	−2.0034	0.134 (0.044–0.388)	<0.001
D dimer	0.0002	1.000 (1.000–1.000)	0.010

Figure 3 Selection of predictive variables using LASSO regression. (A) Characteristics of variable coefficient changes. The horizontal coordinates represent the value of the parameter log(λ), ordinate represents the coefficient of the independent variable. Finally, the coefficients of all independent variables are compressed to 0, and the later the independent variable becomes 0, the greater the contribution to the model. The dashed line on the left indicates the value of the parameter log(λ) when the model error is minimal, the dashed line on the right represents the value of the parameter log(λ) when the model error is amplified by one standard error. (B) Ten-fold cross-validation for adjusting parameters. The horizontal coordinates represent the value of the parameter log(λ), ordinate represents the mean squared error of the model.

Figure 4 Nomogram for predicting whether patients with AP will develop POF. The nomogram created by univariate logistic regression (A). The nomogram created by LASSO (B).

Figure 5 Nomogram performance. ROC curves of the model for predicting POF probabilities in the training cohort (A) and validation cohort (B). Calibration plots for predicting POF probabilities in the training cohort (C) and validation cohort (D). The gray slash represents the ideal reference line, and the black curve represents the predictive performance of the model. The closer the black curve is to the gray slash, the closer the predicted outcome event is to the actual event. Decision curves for the training cohort (E) and validation cohort (F). The abscissa is the threshold probability, and the ordinate is the net benefit rate, when the black dashed line is above the two solid lines, it indicates that the model provides a net benefit.

Figure 6 Dynamic web calculator to predict whether patients with AP will develop POF. The left band inputs variable values, Graphical Summary shows the probability and confidence interval of POF occurrence in the form of pictures.