Development and validation of a nomogram risk prediction model for PICC-related thrombosis in children with hematological malignancies

doi:10.21203/rs.3.rs-4800685/v1

Download PDF

Research Article

Development and validation of a nomogram risk prediction model for PICC-related thrombosis in children with hematological malignancies

https://doi.org/10.21203/rs.3.rs-4800685/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Early recognition and prevention are of great significance in reducing the incidence of Peripheral Intravenous Central Catheter (PICC)-related thrombosis. This study aimed to develop and validate a clinical risk prediction tool for PICC-related thrombosis in children with hematological malignancies.

Methods

Retrospectively selected children with hematological malignancies receiving PICC catheterization from January 2018 to December 2023 in Tongji Hospital as the study subjects and randomly divided into the training and validation sets according to the ratio of 7:3. A total of 54 possible predictor variables were collected from the hospital’s electronic medical record system and subjected to univariate and multivariate analyses. Logistic regression models were used to establish nomograms, which were evaluated by discrimination, calibration degree, and clinical applicability.

Results

519 children were enrolled, of whom 98 (18.9%) were diagnosed with PICC-related thrombosis during retention. The final nomogram model incorporated six independent risk factors, including leukemia, number of catheters, history of catheterization, total parenteral nutrition, post-catheterization D-dimer, and post-catheterization fibrinogen. The area under the receiver operating characteristic curve was 0.844 (95% CI: 0.787 ~ 0.900) and 0.794 (95% CI: 0.698 ~ 0.890) for the training and validation sets, respectively, indicating that the model had good discrimination. All calibration curves showed that the model was well calibration degree. The decision curve analysis showed better net benefit of our model in predicting PICC-related thrombosis risk over a range of threshold probabilities from 5–87% and 91–97% in the training set, and from 4–85% in the validation set.

Conclusions

This nomogram model can be used as an effective tool to predict the risk of PICC-related thrombosis in children with hematological malignancies. It will facilitate pediatricians in early diagnosis, which is critical to reducing the incidence of PICC-related thrombosis.

children

hematological malignancies

PICC-related thrombosis

nomogram

prediction model

Globally, the incidence of hematological malignancy has been increasing since 1990, becoming a major disease that seriously threatens the healthy development of children and adolescents(1–3). Personalized and effective treatment for pediatric patients with hematological malignancies is crucial for improving remission rates and extending survival periods(4, 5). Peripheral Intravenous Central Catheter (PICC) provides patients with a reliable medium and long-term route to intravenous therapy. It is a central venous catheter inserted via peripheral vein puncture with the tip located at the junction of the superior vena cava and the right atrium or at the inferior vena cava above the level of the diaphragm(6). PICC can not only meet the needs of multi-course chemotherapy, hematopoietic stem cell transplantation, parenteral nutrition, blood sampling, and supportive therapy, but also avoid the stimulation and damage of various drugs to blood vessels and tissues, alleviate the pain caused by repeated puncture, and effectively protect peripheral veins(7–9), which has become the first choice of vascular access for hospitalized children with hematological malignancies in China.

Although the PICC has become an essential tool in the treatment and care of children with hematological malignancies, there is a risk of associated complications, including PICC-related thrombosis, phlebitis, catheter migration, catheter obstruction, PICC-associated bloodstream infections(10, 11). Among these, PICC-related thrombosis is one of the most common and serious complications. Children with hematological malignancies often have abnormalities in coagulation and blood routine, which are more likely to contribute to forming a hypercoagulable state in blood and trigger thromboembolic mechanisms when there is an indwelling PICC(12). The incidence of PICC-related thrombosis in children varies significantly by country, age, and healthcare facility, ranging from 4.2–41.5%(13–16). Despite its asymptomatic nature in the early stages, PICC-related thrombosis can still lead to serious side effects, increasing pain and healthcare expenses, causing treatment delays, and impacting prognosis and quality of life. In severe cases, it may culminate in life-threatening pulmonary embolism(17, 18).

The key to reducing the incidence of PICC-related thrombosis is early recognition and prevention. However, there is a lack of child-specific assessment tools in clinical practice. To help healthcare professionals identify high-risk children early and accurately and provide them with individualized prognostic information and preventive strategies, our study aimed to investigate independent risk factors for PICC-related thrombosis in children with hematological malignancies and to develop a predictive model, visualized as a nomogram.

2.1 Study design and participants

We retrospectively collected clinical data of children with hematological malignancies who underwent PICC catheterization between January 2018 and December 2023 at Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology. Inclusion criteria in this study were as follows:(1) age ≤ 18 years old, (2) clinically diagnosed with hematological malignancy, (3) underwent PICC catheterization, and (4) PICC-related thrombosis screening by ultrasound, magnetic resonance, or venography during PICC retention. Exclusion criteria were as follows: (1) PICC catheter placement before admission, (2) PICC retention of less than one day, (3) accidental extubation, and (4) missing values of more than 20% of the data for the included variables. Our study was approved by the Ethics Committee of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology (TJ-IRB20230851). As this study was a retrospective cohort study, informed consent was not required.

2.2 Definition of outcome

The primary outcome indicator was PICC-related thrombosis during hospitalization. PICC-related thrombosis refers to the clumping of red blood cells, platelets, leukocytes, and fibrin in the vein in which the PICC catheter was located, resulting in the formation of catheter thrombus, catheter wall fibrin sheath, and intravascular thrombus(19). If catheter-related deep vein thrombosis (DVT) and/or superficial vein thrombosis (SVT) were detected during PICC retention, the children were considered positive for PICC-related thrombosis and negative if no thrombus was detected.

2.3 Collection of predictor variables

All clinical data were obtained from the electronic medical record system of Tongji Hospital. A total of 54 possible predictive variables were collected in this study, which were classified into patient-related factors, PICC-related factors, treatment-related factors, and laboratory parameters.

Patient-related factors included age, sex, leukemia, history of thrombosis, infection, sepsis, cardiac disease, cardiac arrest, dyspnoea/asphyxia, trauma, renal disease, gastrointestinal disease, hepatobiliary disease, respiratory disease, neurological disease, and autoimmune disease. All of the above disease-related variables were based on clinical diagnosis during hospitalization, including comorbidities and underlying diseases.

PICC-related factors included the location of insertion, catheter size, side of insertion, number of lumens, number of catheters (number of central venous access device including PICC), insertion length (cm), vessel diameter, vein-to-catheter ratio, arm position, catheter displacement, repeated insertions in the same arm, history of catheterization, and catheter indwelling time.

Treatment-related factors include chemotherapy, intensive care unit (ICU) admission, surgery, mechanical ventilation, total parenteral nutrition (TPN), glucocorticoids, vasoactive drugs, blood products, dialysis, hypertonic liquid, furosemide, antibiotics, anticoagulants, cardiac catheterization, and extracorporeal membrane oxygenation (ECMO).

Laboratory parameters included pre-catheterization D-dimer, activated partial thromboplastin time (APTT), fibrinogen (FIB), prothrombin time (PT), platelet count, post-catheterization D-dimer (ug/mL), APTT (s), FIB (g/L), PT (s), and platelet count (*10^9/L). If multiple clinical data were available, the data with the highest value was collected. The above laboratory parameters were converted into categorical variables based on previous study results and clinical boundaries(19, 20).

2.4 Statistical analysis

SPSS 26.0 and R software (V.4.3.2) were used for data analysis. The included children were randomly divided into a training set and a validation set in the ratio of 7:3, which were used for model construction and validation, respectively. Categorical variables were presented as n (%) and continuous variables (normal distribution) as mean ± standard deviation (SD). Univariate analyses were first performed, where t-tests were used to analyze significant differences in continuous variables between groups, while Chi-square tests were used to analyze differences in ratios between groups. Then, variables with p-values less than 0.05 in the univariate analyses were selected for backward elimination (conditional) multivariate logistic regression, and the Akaike’s Information Criteria (AIC) indicator was used to determine the final variables. Based on these variables, a predictive model was established to assess the risk of PICC-related thrombosis in children with hematological malignancies and visualized as a nomogram using the “rms” package for R version 4.3.2.

In this study, the area under the receiver operating characteristic (ROC) curve (AUC) was used to evaluate the discriminant ability of the predictive model, which was low when the AUC ranged from 0.5 to 0.7, medium when it ranged from 0.7 to 0.9, and high when it ranged from 0.9 to 1.0. The Hosmer-Lemeshow test (HL test) was applied as an indicator of model fit, and when the p-value was greater than 0.05, it indicated that there was no very significant difference between the predicted and true values, and the model worked well. In addition, the calibration degree of the model was assessed by calibration curves, and the clinical utility of the model was assessed by decision curve analysis (DCA).

3.1 Baseline characteristics of the study population

The flow chart for the selection of children with hematological malignancies for this retrospective study is shown in Supplementary Material Fig. S1. A total of 519 children were included, of whom 98 (18.9%) were diagnosed with PICC-related thrombosis during catheterization retention, including 364 in the training set, 68 (18.7%) with PICC-related thrombosis, and 155 in the validation set, 30 (19.4%) with PICC-related thrombosis. The mean age of the children in the total population, training set, and validation set was 6.10 ± 3.61, 6.20 ± 3.59, and 5.89 ± 3.67 years, respectively, with 61.3%, 60.2%, and 63.9% males. No statistically significant differences were observed in all basic characteristics between the training and validation sets (Table S1 in the Supplementary Material).

3.2 Model development

Initially, a univariate analysis of all variables in the training set was performed to explore potential risk factors for PICC-related thrombosis (Table 1). The results showed a significant correlation between PICC-related thrombosis in children with hematological malignancies and several predictors, which included sex, leukemia, number of catheters, history of catheterization, ICU admission, TPN, vasoactive drugs, hypertonic liquid, post-catheterization D-dimer, post-catheterization FIB, post-catheterization platelet count (P < 0.05).

Table 1

Univariate analysis of PICC-related thrombosis in training set
Variables	Without thrombosis (N = 296)	With thrombosis (N = 68)	Statistical value	P value
age (years), mean ± SD	6.26 ± 3.56	5.93 ± 3.76	0.683	0.495
sex, n (%)			4.724	0.030
male	186 (62.8)	33 (48.5)
female	110 (37.2)	35 (51.5)
leukemia, n (%)	191 (64.5)	56 (82.4)	8.056	0.005
history of thrombosis, n (%)	2 (0.7)	2 (2.9)	NA*	0.160
infection, n (%)	281 (94.9)	62 (91.2)	0.827^a	0.363
sepsis, n (%)	143 (48.3)	35 (51.5)	0.221	0.638
cardiac disease, n (%)	52 (17.6)	9 (13.2)	0.744	0.388
cardiac arrest, n (%)	2 (0.7)	1 (1.5)	NA*	0.463
dyspnoea/asphyxia, n (%)	22 (7.4)	8 (11.8)	1.372	0.241
trauma, n (%)	4 (1.4)	1 (1.5)	NA*	1.000
renal disease, n (%)	69 (23.3)	18 (26.5)	0.304	0.582
gastrointestinal disease, n (%)	87 (29.4)	25 (36.8)	1.411	0.235
hepatobiliary disease, n (%)	173 (58.4)	34 (50.0)	1.608	0.205
respiratory disease, n (%)	200 (67.6)	39 (57.4)	2.559	0.110
neurological disease, n (%)	28 (9.5)	10 (14.7)	1.628	0.202
autoimmune disease, n (%)	11 (3.7)	7 (10.3)	3.787^a	0.052
location of insertion, n (%)			8.114*	0.106
Basilic vein	255 (86.1)	59 (86.8)
Median cubital vein	2 (0.7)	1 (1.5)
Cephalic vein	16 (5.4)	1 (1.5)
Brachial Vein	21 (7.1)	4 (5.9)
Great saphenous vein	1 (0.3)	1 (1.5)
Femoral vein	1 (0.3)	2 (2.9)
catheter size, n (%)			0.819	0.366
3F	183 (61.8)	38 (55.9)
4F	113 (38.2)	30 (44.1)
side of insertion, n (%)			1.684	0.194
right	250 (84.5)	53 (77.9)
left	46 (15.5)	15 (22.1)
number of lumens, n (%)			-	-
1	296 (100)	68 (100)
>=2	0 (0)	0 (0)
number of catheters, n (%)			27.555	< 0.001
1	275 (92.9)	48 (70.6)
>=2	21 (9.1)	20 (29.4)
insertion length (cm), mean ± SD	26.39 ± 5.36	25.78 ± 5.50	0.836	0.404
vessel diameter (mm), mean ± SD	2.24 ± 0.35	2.24 ± 0.35	−0.131	0.896
vein-to-catheter ratio, n (%)			1.546*	0.375
< 2	95 (32.1)	27 (39.7)
2–3	198 (66.9)	41 (60.3)
>=3	3 (1.0)	0 (0)
arm position, n (%)			0.857	0.651
upper section	40 (13.6)	7 (10.8)
middle section	249 (84.7)	56 (86.2)
lower section	5 (1.7)	2 (3.1)
catheter displacement, n (%)	153 (51.7)	40 (58.8)	1.130	0.288
repeated insertions in the same arm, n (%)	1 (0.3)	2 (2.9)	NA*	0.091
history of catheterization, n (%)	27 (9.1)	22 (32.4)	25.618	< 0.001
catheter indwelling time (day), mean ± SD	180.7 ± 72.8	152.9 ± 115.1	1.909	0.060
chemotherapy, n (%)	285 (96.3)	63 (92.6)	0.982^a	0.322
ICU admission, n (%)	39 (13.2)	18 (26.5)	7.400	0.007
surgery, n (%)	9 (3.0)	3 (4.4)	0.038^a	0.846
mechanical ventilation, n (%)	9 (3.0)	5 (7.4)	1.737^a	0.188
TPN, n (%)	63 (21.3)	31 (45.6)	17.052	< 0.001
glucocorticoids, n (%)	282 (95.3)	62 (91.2)	1.083^a	0.298
vasoactive drugs, n (%)	31 (10.5)	15 (22.1)	6.723	0.010
blood products, n (%)	267 (90.2)	60 (88.2)	0.234	0.628
dialysis, n (%)	2 (0.7)	3 (4.4)	3.273^a	0.070
hypertonic liquid, n (%)	47 (15.9)	19 (27.9)	5.420	0.020
furosemide, n (%)	114 (38.5)	35 (51.5)	3.840	0.050
antibiotics, n (%)	288 (97.3)	63 (92.6)	2.253^a	0.133
anticoagulants, n (%)	27 (9.1)	4 (5.9)	0.745	0.388
cardiac catheterization, n (%)	1 (0.3)	0 (0)	NA*	1.000
ECMO, n (%)	0 (0)	0 (0)	-	-
pre-catheterization D-dimer, n (%)			1.219	0.748
< 0.5	129 (43.6)	25 (36.8)
0.5-1	77 (26.0)	19 (27.9)
1-1.5	37 (12.5)	9 (13.2)
>=1.5	53 (17.9)	15 (22.1)
pre-catheterization APTT, n (%)			1.240	0.265
<=37	170 (57.4)	34 (50.0)
> 37	126 (42.6)	34 (50.0)
pre-catheterization FIB, n (%)			5.603	0.061
< 2	56 (18.9)	8 (11.8)
2–4	191 (64.5)	41 (60.3)
>=4	49 (16.6)	19 (27.9)
pre-catheterization PT, n (%)			NA*	0.160
<=17	294 (99.3)	66 (97.1)
> 17	2 (0.7)	2 (2.9)
pre-catheterization platelet count, n (%)			0.600	0.741
< 125	150 (50.7)	38 (55.9)
125–350	102 (34.5)	21 (30.9)
>=350	44 (14.9)	9 (13.2)
post-catheterization D-dimer, n (%)			57.040	< 0.001
< 0.5	184 (62.2)	15 (22.1)
0.5-1	65 (22.0)	14 (20.6)
1-1.5	28 (9.5)	21 (30.9)
>=1.5	19 (6.4)	18 (26.5)
post-catheterization APTT, n (%)			0.504	0.478
<=37	175 (59.1)	37 (54.4)
> 37	121 (40.9)	31 (45.6)
post-catheterization FIB, n (%)			7.562	0.023
< 2	128 (43.2)	21 (30.9)
2–4	150 (50.7)	37 (54.4)
>=4	18 (6.1)	10 (14.7)
post-catheterization PT, n (%)
<=17	295 (99.7)	66 (97.1)	NA*	0.091
> 17	1 (0.3)	2 (2.9)
post-catheterization platelet count, n (%)			6.662	0.036
< 125	111 (37.5)	36 (52.9)
125–350	138 (46.6)	27 (39.7)
>=350	47 (15.9)	5 (7.4)
* Fisher’s exact test were adopted
^a The continuity correction χ2 test were adopted
Abbreviations: PICC, Peripheral Intravenous Central Catheter; SD, standard deviation; ICU, intensive care unit; TPN, total parenteral nutrition; ECMO, extracorporeal membrane oxygenation; APTT, activated partial thromboplastin time; FIB, fibrinogen; PT, prothrombin time.

Subsequently, these variables that were statistically different in the univariate analyses were analyzed by backward elimination multivariate logistic regression, which utilizes an AIC minima approach to determine the final variable. As the results are shown in Table 2, leukemia, number of catheters, history of catheterization, TPN, post-catheterization D-dimer, and post-catheterization FIB had the strongest correlation with PICC-related thrombosis in children with hematological malignancies.

Table 2

Backward elimination multivariate logistic regression of PICC-related thrombosis
Variables	β coefficient	Standard error	Wald	P-value	OR (95% CI)
leukemia	1.059	0.404	6.883	0.009	2.88 (1.31–6.36)
number of catheters ( > = 2)	1.674	0.443	14.309	< 0.001	5.34 (2.24–12.71)
history of catheterization	1.061	0.405	6.852	0.009	2.89 (1.31–6.40)
TPN	1.098	0.344	10.200	0.001	3.00 (1.53–5.89)
post-catheterization D-dimer
0.5-1	1.059	0.437	5.864	0.015	2.88 (1.22–6.79)
1-1.5	1.940	0.449	18.675	< 0.001	6.96 (2.89–16.78)
>=1.5	2.460	0.497	24.482	< 0.001	11.70 (4.42–31.01)
post-catheterization FIB
2–4	0.558	0.363	2.357	0.125	1.75 (0.86–3.56)
>=4	1.633	0.593	7.589	0.006	5.12 (1.60−16.35)
constant	−4.519	0.568	63.280	< 0.001	0.011
Abbreviations: PICC, Peripheral Intravenous Central Catheter; OR: odds ratio; CI, confidence interval; TPN, total parenteral nutrition; FIB, fibrinogen.

Ultimately, these six independent risk factors were included in the construction of the nomogram model (Fig. 1). Each variable in the nomogram model was scored at different levels according to its degree of influence on the outcome variable. The corresponding probability of PICC-related thrombosis risk can be derived by calculating the total score of all the variables added together and plotting a vertical line on the nomogram.

3.3 Model validation

The prediction model for PICC-related thrombosis in children with hematological malignancies constructed in the training set was evaluated and was also internally validated by using the validation set. In the training set, the AUC value was 0.844 (95% CI: 0.787 ~ 0.900), and the best cut-off value, sensitivity, and specificity were 0.162, 79.4%, and 79.1%, respectively. In the validation set, the AUC value was 0.794 (95% CI: 0.698 ~ 0.890), and the best truncation value, sensitivity, and specificity were 0.127, 86.7%, and 61.6%, respectively (Fig. 2). The difference in AUC between the training set and the internal validation set was 0.05, P = 0.38 for the DeLong test, and there was no statistically significant difference in AUC between the two datasets (P > 0.05). Therefore, the constructed nomogram model has a good discrimination between both the training set and the internal validation set.

The HL test was performed on the training and validation sets, and the p-values were 0.649 and 0.363, indicating that there is no statistically significant difference between the predicted values of the model and the actual observed values. After 1000 bootstrap resampling, the mean absolute errors between the fitting curves and the ideal curves for the training set and the internal validation set were 0.027 and 0.035, respectively (Fig. 3). This suggests good consistency between the curves and good calibration ability of the model to accurately predict the risk probability of PICC-related thrombosis.

The clinical utility of our predictive model was assessed by DCA, as illustrated in Fig. 4. In the curves, the X-axis represents the risk threshold for PICC-related thrombosis and the Y-axis represents the standardized net benefit at different thresholds. The horizontal line labeled "none" indicates that all children had no PICC-related thrombosis, the slash labeled "all" indicates that all children had PICC-related thrombosis, and the blue line represents the decision curve of our clinical prediction model. DCA showed better net benefit of our model in predicting the risk of PICC-related thrombosis over a range of threshold probabilities from 5–87% and 91–97% in the training set, and from 4–85% in the validation set.

PICC has emerged as the primary vascular access for children with hematological malignancies to receive chemotherapy, long-term intravenous fluids, and nutritional support due to its simple operation, good safety, high success rate, long retention time, and low discomfort(21, 22). However, what cannot be ignored is that PICC-related thrombosis has been one of the most common and serious complications. Our results showed that the total incidence of PICC-related thrombosis in children with hematological malignancies was 18.9%, consistent with previous studies(14–16). Early identification of individuals at risk and timely intervention are pivotal in avoiding thrombosis. Currently, however, there is a lack of proprietary assessment tools for children with hematological malignancies. Hence, this study pioneered the development of a nomogram model for predicting PICC-related thrombosis in children with hematological malignancies. The model, which consists of six highly correlated risk factors, showed better predictive performance in both the training and validation sets, with good discrimination, calibration degree, and clinical applicability.

In our study, leukemia was one of the strongest risk factors predicting PICC-related thrombosis in children with hematological malignancies. Leukemia stands as the most prevalent malignancy in children and has a higher incidence of venous thromboembolism (VTE) compared to solid tumors(23, 24). Activation of the coagulation system may be the most important pathogenic factor in leukemia-associated VTE. In the acute leukemic state, abnormally proliferating leukemic cells not only affect the normal hematopoietic function and disturb the internal environment of the child but also contain procoagulants released into the bloodstream(25). In addition, asparaginase, which is the backbone of the chemotherapy regimen for children with leukemia, leads to a decrease in naturally occurring anticoagulants (e.g., protein C, protein S, and antithrombin), and consequently to an increase in thrombin production(26, 27). When child with leukemia has an indwelling PICC, it is more likely to contribute to the further development of a hypercoagulable state of the blood and trigger a thromboembolic mechanism.

PICC catheters are usually made of silicone or special polyurethane that has good tissue compatibility(28). After entering the blood vessels, the temperature of the human body makes it softer and thus less irritating to the vessel. However, this synthetic material is also thought to have the potential to activate the coagulation process due to its lack of an endothelial layer that inhibits platelet coagulation and adhesion(29). The greater the number of catheters, the higher the risk of thrombosis. In addition, multiple catheterizations directly increase the opportunity for vascular intima injury, which also raises the incidence of thrombosis(30). Furthermore, children with multiple catheters were usually more critically ill and needed bed rest for longer periods, which inevitably leads to restricted limb movement, reduced skeletal muscle contraction, muscle relaxation, and weakened vascular support, which leads to slower blood flow and contributes to the formation of PICC-related thrombosis(19, 31).

We observed that the history of catheterization can increase the risk of PICC-related thrombosis in children with hematological malignancies, consistent with the results of Badheka et al.(32) and Shin et al(16). This may be related to venous stenosis caused by catheters through various pathways, including mechanical injury, infection, foreign object stimulation, and abnormal vascular wall damage repair process(33, 34). These factors usually interact with each other and contribute to the development and formation of stenosis. Consequently, when a PICC is placed again in the same limb, it may exacerbate the slow blood flow caused by the venous stenosis and increase the likelihood of blood flow turbulence. This altered hemodynamics can make it easier for blood to form a relatively static area at the stenosis, thus promoting PICC-related thrombosis(35). In addition, stenosis increases the catheter-to-vein ratio, and makes contact and friction between the PICC and the vessel wall more frequent, which may cause damage and further activate the process of clot formation(14).

TPN refers to the provision of all daily nutrient requirements by intravenous infusion. Previous studies have shown a significantly higher incidence of thrombosis in children with intestinal disorders who were placed on central venous catheters for long-term TPN(36). In our study, most of the children requiring TPN did not have concomitant intestinal disease and needed only short- to medium-term TPN. Nevertheless, the results of this study suggest that TPN remains a significant risk factor for PICC-related thrombosis. Firstly, TPN mixtures generally contain high concentrations of glucose, amino acids, and fat emulsion, which may alter plasma osmolality and increase blood viscosity, thereby increasing resistance to blood flow. Additionally, components such as glucose in TPN mixtures may favor the procoagulant state of monocytes(37). Moreover, calcium, one of the major ions in the TPN mixture, is an essential activator of the coagulation cascade, and its supplementation promotes activation of the coagulation cascade and increases the risk of thrombosis(38).

FIB is a protein with coagulant function involved in the body’s coagulation and hemostatic processes. As the most abundant coagulation factor in plasma, the normal levels range from 1.5 to 3.5 g/L(39). D-dimer is a soluble fibrin degradation product derived from plasmin-mediated degradation of cross-linked fibrin. Its elevated concentration represents activation of the coagulation system and secondary hyperfibrinolysis in the body(40). Both of these indicators can reflect the hypercoagulable state of blood. The results of this study showed that FIB concentration and D-dimer concentration after catheterization were the two most critical risk factors for PICC-related thrombosis in children with hematological malignancies, and children with high concentrations were more likely to develop thrombosis. Therefore, after PICC placement, healthcare professionals should routinely screen children for FIB and D-dimer concentrations. Children with high or persistently elevated levels of both indicators should be alerted to the development of PICC-related thrombosis and given appropriate prophylactic anticoagulant therapy. It is worth noting that even though FIB and D-dimer have certain applications in the diagnosis of thrombosis(41–43), they cannot be the only indicators to confirm the diagnosis of thrombosis. The risk of thrombosis should be assessed in conjunction with other key risk factors.

In this study, we constructed the first PICC-related thrombosis prediction model for children with hematological malignancies and visualized the risk prediction results of the model using a nomogram, which has good discrimination, calibration degree, and clinical applicability. The model can be used as an effective assessment tool for PICC-related thrombosis in children with hematological malignancies, providing useful predictive information for clinical staff and parents of children. Through focused monitoring and active intervention in high-risk children, the occurrence of PICC-related thrombosis can be effectively reduced, and unnecessary complications and treatment delays can be avoided. This will simultaneously help improve the efficiency of healthcare resource utilization and reduce the burden on patients and the healthcare system.

Nevertheless, this study also has some limitations. (1) The construction of the model was based on retrospective data, which may have selection bias, missing data, and incompleteness. (2) The sample size was relatively limited and all were from the same medical center, which may affect the stability of the model and require caution in interpreting the findings. (3) When establishing the model, multiple continuous variables were converted to categorical variables for clinical convenience, which to some extent resulted in a loss of information and may affect the accuracy of the model. (4) Lack of external validation makes it impossible to understand the generalization performance of the model on new datasets. (5) The number of factors included in our study was limited, and the relationship between other relevant factors and the risk of PICC-related thrombosis is still unclear. Therefore, future studies with larger sample sizes of children with hematological malignancies are needed to determine whether factors other than those already in the model are independently associated with PICC-related thrombosis to further optimize and update our model.

In conclusion, we identified several of the most critical risk factors for PICC-related thrombosis in children with hematological malignancies, including leukemia, number of catheters, history of catheterization, TPN, post-catheterization D-dimer, post-catheterization FIB, and constructed a nomogram prediction model with good effect. This nomogram model can be used as an effective tool to predict the risk of PICC-related thrombosis in children with hematological malignancies, which can help clinical staff identify high-risk children early and adopt suitable thromboprophylaxis strategies to reduce thrombosis and improve clinical decision-making and treatment safety.

PICC

Peripheral Intravenous Central Catheter

DVT

deep vein thrombosis

SVT

superficial vein thrombosis

ICU

intensive care unit

TPN

total parenteral nutrition

ECMO

extracorporeal membrane oxygenation

APTT

activated partial thromboplastin time

FIB

fibrinogen

prothrombin time

standard deviation

AIC

Akaike’s Information Criteria

ROC

receiver operating characteristic

AUC

area under ROC

HL test

Hosmer-Lemeshow test

DCA

decision curve analysis

VTE

venous thromboembolism.

Ethics approval and consent to participate:
This study was performed in line with the principles of the Declaration of Helsinki. The institutional review board of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology approved this study (TJ-IRB20230851). As this study was a retrospective cohort study, informed consent was not required.

Consent for publication:
Not applicable

Competing interests:

The authors declare that they have no competing interests.

Funding:

This study was supported by the Fundamental Research Funds for the Central Universities (grant number:YCJJ20230244) and Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology Research Fund (grant number:2022C09).

Authors’ contributions:

Maoling Fu: Conceptualization, Methodology, Data curation, Writing – original draft, Funding acquisition. Qiaoyue Yang: Data curation, Software. Quan Yuan: Formal analysis, Data curation, Writing – original draft. Xiao Wu: Software, Writing – review & editing. Ting Yang: Conceptualization, Writing – review & editing. Xinyu Li: Formal analysis, Data curation, Writing – original draft. Lexue Jiang: Writing – review & editing, Formal analysis. Xiuli Qin: Writing – review & editing, Data curation. Huiping Yan: Writing – review & editing. Genzhen Yu: Software, Supervision, Writing – review & editing, Funding acquisition.

Acknowledgements:

Not applicable

Availability of data and materials:

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Zhang N, Wu J, Wang Q, Liang Y, Li X, Chen G, et al. Global burden of hematologic malignancies and evolution patterns over the past 30 years. Blood Cancer J. 2023;13(1):82. https://doi.org/10.1038/s41408-023-00853-3.
Ni X, Li Z, Li X, Zhang X, Bai G, Liu Y, et al. Socioeconomic inequalities in cancer incidence and access to health services among children and adolescents in China: a cross-sectional study. Lancet. 2022;400(10357):1020–32. https://doi.org/10.1016/s0140-6736(22)01541-0.
Lam CG, Howard SC, Bouffet E, Pritchard-Jones K. Science and health for all children with cancer. Science. 2019;363(6432):1182–6. https://doi.org/10.1126/science.aaw4892.
Shah RM. Contemporary haploidentical stem cell transplant strategies in children with hematological malignancies. Bone Marrow Transpl. 2021;56(7):1518–34. https://doi.org/10.1038/s41409-021-01246-5.
Hemminki K, Hemminki J, Försti A, Sud A. Survival in hematological malignancies in the Nordic countries through a half century with correlation to treatment. Leukemia. 2023;37(4):854–63. https://doi.org/10.1038/s41375-023-01852-w.
Gorski LA, Hadaway L, Hagle ME, Broadhurst D, Clare S, Kleidon T et al. Infusion Therapy Standards of Practice, 8th Edition. J Infus Nurs. 2021;44(1S Suppl 1):S1-s224. https://doi.org/10.1097/nan.0000000000000396
Mielke D, Wittig A, Teichgräber U. Peripherally inserted central venous catheter (PICC) in outpatient and inpatient oncological treatment. Support Care Cancer. 2020;28(10):4753–60. https://doi.org/10.1007/s00520-019-05276-0.
Yamaguchi RS, Noritomi DT, Degaspare NV, Muñoz GOC, Porto APM, Costa SF, et al. Peripherally inserted central catheters are associated with lower risk of bloodstream infection compared with central venous catheters in paediatric intensive care patients: a propensity-adjusted analysis. Intensive Care Med. 2017;43(8):1097–104. https://doi.org/10.1007/s00134-017-4852-7.
Pitiriga V, Bakalis J, Theodoridou K, Kanellopoulos P, Saroglou G, Tsakris A. Lower risk of bloodstream infections for peripherally inserted central catheters compared to central venous catheters in critically ill patients. Antimicrob Resist Infect Control. 2022;11(1):137. https://doi.org/10.1186/s13756-022-01180-1.
Wang P, Soh KL, Ying Y, Liu Y, Huang X, Huang J. Risk of VTE associated with PORTs and PICCs in cancer patients: A systematic review and meta-analysis. Thromb Res. 2022;213:34–42. https://doi.org/10.1016/j.thromres.2022.02.024.
Pitiriga V, Bakalis J, Theodoridou K, Dimitroulia E, Saroglou G, Tsakris A. Comparison of microbial colonization rates between central venous catheters and peripherally inserted central catheters. Antimicrob Resist Infect Control. 2023;12(1):74. https://doi.org/10.1186/s13756-023-01285-1.
Athale UH, Chan AK. Thromboembolic complications in pediatric hematologic malignancies. Semin Thromb Hemost. 2007;33(4):416–26. https://doi.org/10.1055/s-2007-976177.
Gnannt R, Waespe N, Temple M, Amirabadi A, Liu K, Brandão LR, et al. Increased risk of symptomatic upper-extremity venous thrombosis with multiple peripherally inserted central catheter insertions in pediatric patients. Pediatr Radiol. 2018;48(7):1013–20. https://doi.org/10.1007/s00247-018-4096-x.
Menéndez JJ, Verdú C, Calderón B, Gómez-Zamora A, Schüffelmann C, de la Cruz JJ, et al. Incidence and risk factors of superficial and deep vein thrombosis associated with peripherally inserted central catheters in children. J Thromb Haemost. 2016;14(11):2158–68. https://doi.org/10.1111/jth.13478.
Dubois J, Rypens F, Garel L, David M, Lacroix J, Gauvin F. Incidence of deep vein thrombosis related to peripherally inserted central catheters in children and adolescents. CMAJ. 2007;177(10):1185–90. https://doi.org/10.1503/cmaj.070316.
Shin HS, Towbin AJ, Zhang B, Johnson ND, Goldstein SL. Venous thrombosis and stenosis after peripherally inserted central catheter placement in children. Pediatr Radiol. 2017;47(12):1670–5. https://doi.org/10.1007/s00247-017-3915-9.
Patel N, Petersen TL, Simpson PM, Feng M, Hanson SJ. Rates of Venous Thromboembolism and Central Line-Associated Bloodstream Infections Among Types of Central Venous Access Devices in Critically Ill Children. Crit Care Med. 2020;48(9):1340–8. https://doi.org/10.1097/ccm.0000000000004461.
Ullman AJ, Gibson V, Takashima MD, Kleidon TM, Schults J, Saiyed M, et al. Pediatric central venous access devices: practice, performance, and costs. Pediatr Res. 2022;92(5):1381–90. https://doi.org/10.1038/s41390-022-01977-1.
Tian L, Su Y, Gao H, Wang L, Zeng J, Yang Q, et al. Development and validation of a nomogram model for central venous access device-related thrombosis in hospitalized children. Eur J Pediatr. 2023;182(11):4909–19. https://doi.org/10.1007/s00431-023-05078-x.
Li S, Luo Y, Deng J, Zeng J, Fan M, Wang T, et al. Risk factors for central venous catheter-related thrombosis in hospitalized children: a single-center a retrospective cohort study. Transl Pediatr. 2022;11(11):1840–51. https://doi.org/10.21037/tp-22-529.
Chen L, Lu Y, Wang L, Pan Y, Zhou X. Construction of a nomogram risk prediction model for PICC-related venous thrombosis and its application. Asian J Surg. 2024;47(1):107–11. https://doi.org/10.1016/j.asjsur.2023.05.043.
Lin Y, Zeng Z, Lin R, Zheng J, Liu S, Gao X. The Caprini thrombosis risk model predicts the risk of peripherally inserted central catheter-related upper extremity venous thrombosis in patients with cancer. J Vasc Surg Venous Lymphat Disord. 2021;9(5):1151–8. https://doi.org/10.1016/j.jvsv.2020.12.075.
Gade IL, Brækkan S, Næss IA, Hansen JB, Rosendaal F, Cannegieter S, et al. Epidemiology of venous thromboembolism in hematological cancers: The Scandinavian Thrombosis and Cancer (STAC) cohort. Thromb Res. 2017;158:157–60. https://doi.org/10.1016/j.thromres.2017.09.002.
Kekre N, Connors JM. Venous thromboembolism incidence in hematologic malignancies. Blood Rev. 2019;33:24–32. https://doi.org/10.1016/j.blre.2018.06.002.
Ziegler S, Sperr WR, Knöbl P, Lehr S, Weltermann A, Jäger U, et al. Symptomatic venous thromboembolism in acute leukemia. Incidence, risk factors, and impact on prognosis. Thromb Res. 2005;115(1–2):59–64. https://doi.org/10.1016/j.thromres.2004.07.016.
Nowak-Göttl U, Heinecke A, von Kries R, Nürnberger W, Münchow N, Junker R. Thrombotic events revisited in children with acute lymphoblastic leukemia: impact of concomitant Escherichia coli asparaginase/prednisone administration. Thromb Res. 2001;103(3):165–72. https://doi.org/10.1016/s0049-3848(01)00286-9.
Levy-Mendelovich S, Barg AA, Kenet G. Thrombosis in pediatric patients with leukemia. Thromb Res. 2018;164(Suppl 1):S94–7. https://doi.org/10.1016/j.thromres.2018.01.019.
Ding N, Peng H, Zhao W, Yi Y, Ma Y, Guo Y, et al. Effects of Peripherally inserted Central Catheter (PICC) materials and designs on reduction of PICC-related complications: A systematic review and meta-analysis. Int Wound J. 2023;21(3):e14468. https://doi.org/10.1111/iwj.14468.
Ekdahl KN, Lambris JD, Elwing H, Ricklin D, Nilsson PH, Teramura Y, et al. Innate immunity activation on biomaterial surfaces: a mechanistic model and coping strategies. Adv Drug Deliv Rev. 2011;63(12):1042–50. https://doi.org/10.1016/j.addr.2011.06.012.
Song X, Lu H, Chen F, Bao Z, Li S, Li S, et al. A longitudinal observational retrospective study on risk factors and predictive model of PICC associated thrombosis in cancer patients. Sci Rep. 2020;10(1):10090. https://doi.org/10.1038/s41598-020-67038-x.
Wang C, Cui F, Li J, Yuan X, Wang J, Liu L, et al. Risk Factors for Venous Thromboembolism in Hospitalized Patients in the Chinese Population. Open Life Sci. 2018;13:82–9. https://doi.org/10.1515/biol-2018-0012.
Badheka AV, Hodge D, Ramesh S, Bloxham J, Espinoza E, Allareddy V, et al. Catheter related thrombosis in hospitalized infants: A neural network approach to predict risk factors. Thromb Res. 2021;200:34–40. https://doi.org/10.1016/j.thromres.2021.01.009.
Sohail MA, Vachharajani TJ, Anvari E. Central Venous Catheters for Hemodialysis-the Myth and the Evidence. Kidney Int Rep. 2021;6(12):2958–68. https://doi.org/10.1016/j.ekir.2021.09.009.
Ge X, Cavallazzi R, Li C, Pan SM, Wang YW, Wang FL. Central venous access sites for the prevention of venous thrombosis, stenosis and infection. Cochrane Database Syst Rev. 2012;2012(3):Cd004084. https://doi.org/10.1002/14651858.CD004084.pub3.
Nesbitt WS, Mangin P, Salem HH, Jackson SP. The impact of blood rheology on the molecular and cellular events underlying arterial thrombosis. J Mol Med (Berl). 2006;84(12):989–95. https://doi.org/10.1007/s00109-006-0101-1.
Smitherman AB, Alexander T, Connelly M, Snavely AC, Weston BW, Liles EA, et al. The incidence of catheter-associated venous thrombosis in noncritically ill children. Hosp Pediatr. 2015;5(2):59–66. https://doi.org/10.1542/hpeds.2014-0041.
Tian L, Li W, Su Y, Gao H, Yang Q, Lin P, et al. Risk Factors for Central Venous Access Device-Related Thrombosis in Hospitalized Children: A Systematic Review and Meta-Analysis. Thromb Haemost. 2021;121(5):625–40. https://doi.org/10.1055/s-0040-1720976.
Kakzanov V, Monagle P, Chan AK. Thromboembolism in infants and children with gastrointestinal failure receiving long-term parenteral nutrition. JPEN J Parenter Enter Nutr. 2008;32(1):88–93. https://doi.org/10.1177/014860710803200188.
de Moerloose P, Boehlen F, Neerman-Arbez M. Fibrinogen and the risk of thrombosis. Semin Thromb Hemost. 2010;36(1):7–17. https://doi.org/10.1055/s-0030-1248720.
Johnson ED, Schell JC, Rodgers GM. The D-dimer assay. Am J Hematol. 2019;94(7):833–9. https://doi.org/10.1002/ajh.25482.
Sartori M, Migliaccio L, Favaretto E, Cini M, Legnani C, Palareti G, et al. D-dimer for the diagnosis of upper extremity deep and superficial venous thrombosis. Thromb Res. 2015;135(4):673–8. https://doi.org/10.1016/j.thromres.2015.02.007.
Cosmi B, Legnani C, Cini M, Tomba S, Migliaccio L, Borgese L, et al. Thrombotic burden, D-dimer levels and complete compression ultrasound for diagnosis of acute symptomatic deep vein thrombosis of the lower limbs. Thromb Res. 2022;213:163–9. https://doi.org/10.1016/j.thromres.2022.03.019.
Chandrashekar A, Singh G, Jonah G, Sikalas N, Labropoulos N. Mechanical and Biochemical Role of Fibrin Within a Venous Thrombus. Eur J Vasc Endovasc Surg. 2018;55(3):417–24. https://doi.org/10.1016/j.ejvs.2017.12.002.

renamed315fb.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

Development and validation of a nomogram risk prediction model for PICC-related thrombosis in children with hematological malignancies

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

1. Background

2. Methods

2.1 Study design and participants

2.2 Definition of outcome

2.3 Collection of predictor variables

2.4 Statistical analysis

3. Results

3.1 Baseline characteristics of the study population

3.2 Model development

3.3 Model validation

4. Discussion

5. Conclusions

Abbreviations

Declarations

Consent for publication:
Not applicable

Competing interests:

Funding:

Authors’ contributions:

Acknowledgements:

Availability of data and materials:

References

Supplementary Files

Status:

Version 1

Development and validation of a nomogram risk prediction model for PICC-related thrombosis in children with hematological malignancies

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

1. Background

2. Methods

2.1 Study design and participants

2.2 Definition of outcome

2.3 Collection of predictor variables

2.4 Statistical analysis

3. Results

3.1 Baseline characteristics of the study population

3.2 Model development

3.3 Model validation

4. Discussion

5. Conclusions

Abbreviations

Declarations

Consent for publication: Not applicable

Competing interests:

Funding:

Authors’ contributions:

Acknowledgements:

Availability of data and materials:

References

Supplementary Files

Status:

Version 1

Consent for publication:
Not applicable