Abstract
Data confirming parenteral nutrition (PN) as an independent risk factor for catheter-related infection (CRI) are scarce and not recent. This study aims to estimate the effect of PN on CRI in cancer patients after catheter placement. A retrospective cohort analysis of the French National Health Data System was conducted. A total of 5300 gastric, 5975 pancreatic, and 15,102 lung cancer patients with central venous catheter insertion between 2012 and 2016 were included. Potential confounders for CRI were evaluated using Poisson regression. Incidence rates (IRs) of CRI were calculated. Incidence rate ratio (IRR) was estimated to assess the effect of exposure to PN on infection occurrence in all patients and among patients receiving chemotherapy (CT). The independent risk factors found to be associated with CRI were age, sex, Charlson Comorbidity Index, PN, CT, metastasis, and history of infection. After adjustment, there was an increased risk of CRI in patients with versus without PN in gastric (IRR: 1.1165; 95% confidence interval [95% CI]: 1.0119–1.2319), pancreatic (IRR: 1.2071; 95% CI: 1.0985–1.3265), and lung cancer (IRR: 1.1850; 95% CI: 1.0832–1.2962). In patients who received PN, IRs of CRI were higher than in those who did not, for gastric (34.84% vs 29.69%, respectively; P = .0001), pancreatic (38.89% vs 28.58%, respectively; P < .0001), and lung (23.64% vs 22.13%, respectively; P = .0967) cancer cohorts. The results of this National administrative claims database analysis suggested that PN might be one of the independent risk factors for CRI. However, the risk was smaller than previously reported, regardless of whether the patient was receiving CT or not.
Keywords: cancer, catheter-related infection, chemotherapy, parenteral nutrition
1. Introduction
Cancer patients often require central venous catheters (CVCs) for cancer treatment, blood transfusions, and parenteral nutrition (PN). The use of PN in cancer patients improves the weight and energy intake in this population and has a positive impact on clinical outcomes. However, the administration of PN through CVC may increase the risk of catheter-related infection (CRI) leading to significant morbidity and mortality.[1] As such, catheter-related sepsis has been reported to occur in 1.3% to 26.2% of the catheters used to administer PN.[2]
Yet, clinical data confirming the safety of PN administration in cancer patients is still the subject of much controversy. First, patients with cancer are particularly susceptible to infection, as they are often immune-compromised due to cancer, anti-cancer treatments, age, or presence of comorbidities. Thus, data suggesting that PN as an independent risk factor for CVC-related infection after adjusting for factors such as age, severity of disease, cancer characteristics, chemotherapy (CT) side effects, or weight loss are insufficient. Moreover, in studies performed more than 20 years ago,[1,2] higher infection rates were observed among patients using PN compared with either no nutrition or enteral nutrition. However, several recent studies conducted during the past 10 years have failed to support these outdated results.[3–5] Potential reasons for such discordant results include improved glycemic control, avoidance of overfeeding, and improved sterility and CVC care. In France a prospective study published in 2013 confirmed that PN was an independent risk factor for CVC-related infection in digestive cancer patients who underwent therapy. However, this was a monocentric study on 425 subjects with some limitations.[6]
In France, a national surveillance program is available to pilot the strategy of prevention of infections linked to central catheters (Surveillance and Prevention of Infections Associated with Invasive Devices).[7] Prevalence and incidence of CVC-related infections are reported annually, as well as a description for the types of responsible microorganisms. However, no specific data about the use of PN or its association with these infections are available.
This study aimed to estimate the incidence of CRI in patients with gastrointestinal (GI) (gastric and pancreatic) cancers and 1 non-GI (lung) cancer, receiving PN or not, using a population-level national health care claims database in France.
2. Methods
2.1. Study design
A retrospective cohort analysis of the French National Health Data System (SNDS) administrative claims database was conducted to determine the incidence of CRI in cancer patients receiving PN and those not receiving PN.
Patients were followed from the date of their catheter placement (index date) until onset date of infection event or other censoring date, such as the date of death, or end of study (December 2018), whichever occurred first. Patients included in each of the 3 cancer cohorts were assigned to 1 of 2 groups depending on their exposure or no to PN after catheter placement. Outcome event included first infection that occurred during a hospitalization after catheter placement (Fig. 1). The study reporting was performed in accordance with the international STROBE reporting standards for retrospective cohort analysis.[8] Data access and running of the analysis would have not been made possible without facilitation and approval by the French data protection authority and the French Ethical and Scientific Committee for Research, Studies and Evaluations in the Health Field. Approval number TPS 704814 was granted on September 11, 2019. French Ethical and Scientific Committee for Research, Studies and Evaluations in the Health Field is a governmental organization responsible for issuing, before authorization from the French data protection authority, opinions on study projects requiring the use of public health data in the French administrative health care database (SNDS).
Figure 1.
Study scheme, PN: parenteral nutrition.
2.2. Data source
The SNDS claims database is representative of 99% of the French population.[9] It links data from the national health insurance database (SNIIRAM), data from the National hospital discharge database, and data from the national death registry (Centre for Epidemiology on Medical Causes of Death/Database on Medical Causes of Death in France, run by Centre for Epidemiology on Medical Causes of Death) using pseudonymization of the French unique national identifier.[10] The database includes demographic data, all the healthcare events presented for reimbursement for more than 66 million individuals (medical, pharmacy, and procedure claims). It also contains the long-term chronic disease (Affection de Longue Durée) status, allowing for a full medical reimbursement. In-hospital diagnoses are coded according to the International Classification of Diseases, 10th revision (ICD-10)[11]; and are classified as a primary diagnosis (main reason for admission); a related diagnosis (that specifies the disease context of the primary diagnosis); or associated diagnosis related to other comorbidities. Medical procedures were encoded according to the Classification Commune des Actes Médicaux.[12] Drugs dispensed during the hospital stay and outpatient pharmacies are coded according to the Common Dispensing Units classification or Presentation Identifier Code.[13] Medical devices and medical nutrition present in the list of products and services were identified via the Liste des produits et prestations codes.[14]
Access to the SNDS database was granted after submission and approval by specific government bodies and national ethical committee, in strict compliance to the local regulations.
2.3. Study population
Three types of cancer were considered for this study, 2 GI cancers: gastric and pancreatic, and 1 non-GI cancer: lung. Patients in the database between January 2012 and December 2016 aged ≥ 18 years, with at least 1 claim for the selected cancer, and with no history of diagnosis of any type of cancers within 2 years before enrollment were included in the initial data cut. The cancer diagnosis was identified by ICD-10 codes either during a hospitalization (hospital admission or homecare), or by a cancer diagnosis used for patient enrollment in an Affection de Longue Durée. Of these patients, those who had at least 1 claim for CVC insertion were included in the study.
2.4. Outcomes
2.4.1. Primary outcome
The primary outcome was the incidence of CRI in cancer patients with central catheter. Outcome event included first CRI that occurred during a hospitalization after catheter placement. CRIs were detected through hospitalizations only (given the paucity of diagnostic codes in the outpatient setting of care in the SNDS database) and identified using ICD-10 diagnostic codes consistent with infections associated with medical device, microorganisms responsible for bacteria associated with a vascular device, or sepsis. Since infections are recurrent, patients with history of infection were not excluded. The occurrence of each type of event during follow-up was also described.
2.4.2. Secondary outcomes
Baseline variables and those known to be associated with CRIs were described. Demographic data included age at cancer diagnosis (with categories of 18–49, 50–59, 60–69, and ≥ 70 years), age at catheter placement, and sex. Comorbidities within a 12-month pre-catheter placement period were assessed using the Charlson Comorbidity Index (CCI) score calculated on a range of seventeen conditions measuring a patient’s overall comorbidity burden.[15] The CCI adapted to the French National Health Insurance database, as defined by Bannay et al,[16] was used in this study. The presence of moderate or severe malnutrition within 12-month before catheter placement was identified using ICD-10 codes for diagnoses. The presence of metastasis at baseline was defined as the presence of any code for secondary neoplasm or the presence of a diagnosis of a second cancer other than the primary 1 during the period between cancer diagnosis and catheter placement. CT treatment at baseline was identified by hospital admissions for CT using disease-related group codes or ICD-10 code Z51.1 within 2 months before or after catheter placement. The use of PN in the database was identified in the inpatient and outpatient settings through a clinical procedure for PN (Classification Commune des Actes Médicaux codes), introduction of a related medical device related to PN (Liste des produits et prestations codes), or a pharmacy claim for PN product delivery (Presentation Identifier Code codes). History of infection was defined as the presence of an infection event (any type) between cancer diagnosis and catheter placement.
2.5. Statistical analysis
All analyses were performed using SAS statistical software (Version 9.4, SAS Institute Inc). Analysis was independently performed for each of the 3 cancers.
Categorical variables were shown as numbers (%), and continuous parameters were shown as means (standard deviations) or medians (interquartile ranges). IRs of CRI per patient-month were calculated as the total number of patients with infection divided by the total time at risk after catheter placement. IRR and 95% confidence interval (95% CI) were estimated using Poisson regression to assess the effect of exposure to PN on infection occurrence. Adjusted IRR was also estimated in a multivariable model after adjustment for potential confounders of infection selected according to their clinical relevance, including age at catheter placement categorized as (<50, 50–59, 60–69, ≥70 years), sex, CCI categorized as (0, 1, 2, 3–4, ≥5), history of infection, metastasis, and treatment with CT. P-values < .05 were considered as statistically significant. IRR were also estimated among a subset of patients receiving CT.
3. Results
3.1. Description of dataset
A total of 26,377 cancer patients were included in the study, of which 5300 (20.1%) patients were diagnosed with gastric cancer; 5975 (22.7%) patients with pancreatic cancer, and 15,102 (57.3%) patients with lung cancer.
Number of patients included per year from 2012 to 2016 are available in Figure S1, Supplemental Digital Content, https://links.lww.com/MD/P126.
3.2. Demographic and clinical characteristics of study sample
Patients’ characteristics by disease site are shown in Table 1. Most patients were older than 60 years at the time of cancer diagnosis, male and with metastatic disease at the time of CVC insertion. Baseline characteristics of eligible patients in each cohort were also examined and stratified by exposure status (Table S1, Supplemental Digital Content, https://links.lww.com/MD/P127).
Table 1.
Demographic and clinical characteristics of study sample.
| Gastric cancer N = 5,300 | Pancreatic cancer N = 5,975 |
Lung cancer N = 15,102 |
|
|---|---|---|---|
| Age at cancer diagnosis, mean [SD] | 69.16 [13.44] | 68.17 [11.89] | 64.36 [10.71] |
| Age groups at cancer diagnosis, n (%) | |||
| <50 years | 474 (8.94%) | 407 (6.81%) | 1246 (8.25%) |
| 50–59 years | 710 (13.40%) | 933 (15.62%) | 3586 (23.75%) |
| 60–69 years | 1346 (25.40%) | 1800 (30.13%) | 5564 (36.84%) |
| ≥70 years | 2770 (52.26%) | 2835 (47.45%) | 4706 (31.16%) |
| Age at catheter placement, mean [SD] | 69.28 [13.42] | 68.26 [11.87] | 64.55 [10.71] |
| Male, n (%) | 3377 (63.72%) | 3284 (54.96%) | 10897 (72.16%) |
| Time from cancer diagnosis to catheter placement (months), mean [SD] | 4.08 [6.66] | 2.97 [6.07] | 4.95 [8.15] |
| Presence of malnutrition, n (%) | 2532 (47.77%) | 2846 (47.63%) | 4903 (32.47%) |
| Metastasis, n (%) | 2977 (56.17%) | 3261 (54.58%) | 9850 (65.22%) |
| Chemotherapy at baseline, n (%) | 2072 (39.09%) | 2391 (40.02%) | 7021 (46.49%) |
| History of infection, n (%) | 2518 (47.51%) | 2958 (49.51%) | 6755 (44.73%) |
| CCI, mean [SD] | 1.60 [1.82] | 1.66 [1.83] | 2.01 [1.92] |
| Comorbidities, n (%) | |||
| Myocardial infarction | 526 (9.92%) | 466 (7.80%) | 1895 (12.55%) |
| Congestive heart failure | 569 (10.74%) | 495 (8.28%) | 2186 (14.47%) |
| Peripheral vascular disease | 1298 (24.49%) | 1378 (23.06%) | 4672 (30.94%) |
| Cerebrovascular disease | 385 (7.26%) | 352 (5.89%) | 1518 (10.05%) |
| Dementia | 167 (3.15%) | 159 (2.66%) | 270 (1.79%) |
| Chronic pulmonary disease | 899 (16.96%) | 842 (14.09%) | 6884 (45.58%) |
| Connective tissue disease | 78 (1.47%) | 80 (1.34%) | 307 (2.03%) |
| Ulcer disease | 819 (15.45%) | 318 (5.32%) | 363 (2.40%) |
| Mild liver disease | 202 (3.81%) | 312 (5.22%) | 625 (4.14%) |
| Diabetes without end organ damage | 621 (11.72%) | 1349 (22.58%) | 1419 (9.40%) |
| Hemiplegia | 133 (2.51%) | 135 (2.26%) | 1068 (7.07%) |
| Moderate or severe renal disease | 574 (10.83%) | 628 (10.51%) | 1727 (11.44%) |
| Diabetes with end organ damage | 554 (10.45%) | 803 (13.44%) | 1569 (10.39%) |
| Any tumor (non-metastatic) | – | – | – |
| Moderate or severe liver disease | 100 (1.89%) | 312 (5.22%) | 293 (1.94%) |
| Metastatic solid tumor | – | – | – |
| HIV/AIDS | 17 (0.32%) | 19 (0.32%) | 96 (0.64%) |
AIDS = acquired immunodeficiency syndrome, CCI = Charlson Comorbidity Index, HIV = human immunodeficiency viruses, SD = standard deviation.
3.3. Incidence of CRI
PN use was the highest in gastric patients (40.5%) followed by pancreatic cancer (30.7%) and lung (16.7%). Incidences of CRI observed in the 3 cohorts were 31.8%, 31.8%, 22.4% in gastric, pancreatic, and lung cancer patients respectively, whereas incidences of CRI were 15.4%, 14.5%, and 11.6% in gastric, pancreatic, and lung patients with no history of infection respectively (Table 2).
Table 2.
Outcomes occurring during the observation period.
| Gastric cancer N = 5,300 |
Pancreatic cancer N = 5,975 |
Lung cancer N = 15,102 |
|
|---|---|---|---|
| Observation period (month) | |||
| Mean [SD] | 20.09 [23.84] | 16.52 [22.12] | 15.54 [22.28] |
| Median (Q1-Q3) | 7 (1.18–35.37) | 4 (0.82–27.73) | 3 (0.49–26.64) |
| Min-Max | 0.03–88.31 | 0.03–83.71 | 0.03–88.74 |
| Use of PN, n (%) | 2144 (40.45%) | 1836 (30.73%) | 2525 (16.72%) |
| Death, n (%) | 2241 (42.28%) | 2876 (48.13%) | 8541 (56.56%) |
| Incidence of CRI, n (%) | 1684 (31.77%) | 1897 (31.75%) | 3380 (22.38%) |
| Incidence of CRI among patients with no history of infection, n (%) | 816 (15.40%) | 866 (14.49%) | 1744 (11.55%) |
| Chemotherapy during follow-up, n (%) | 2389 (45.08%) | 2932 (49.07%) | 7682 (50.87%) |
| Chemotherapy before catheter placement, n (%) | 691 (13.04% of all patients) 691 (41.03% of patients with infection) |
993 (16.62% of all patients) 993 (52.35% of patients with infection) |
1711 (11.33% of all patients) 1711 (50.62% of patients with infection) |
CRI = catheter-related infection, PN = parenteral nutrition, Q1 = first quartile, Q3 = third quartile, SD = standard deviation.
In the gastric and pancreatic cancer cohorts there were significant differences in the IRs CRI between patients who received PN (34.84% and 38.89%, respectively) and those who did not (29.69% and 28.58%, respectively); P = .0001 and P < .0001, respectively. Whereas in the lung cancer cohort there were no differences in the IRs CRI between patients who received PN and those who did not (23.64% vs 22.13%, respectively; P = .0967).
In the multivariate analysis, the independent risk factors found to be associated with CRIs were age, sex, CCI score, PN, CT, metastasis, and history of infection (Table 3). After adjusting for these confounding variables, use of PN was associated with an increased risk of CRI in all cancer cohorts with an estimated increase in risk of 11.7% (IRR = 1.1165; 95% CI: 1.0119–1.2319; P = .0281) in the gastric; 20.7% (IRR = 1.2071; 95% CI: 1.0985–1.3265; P < .0001) in the pancreatic; and 18.5% (IRR = 1.1850; 95% CI: 1.0832–1.2962; P = .0002) in the lung cancer cohorts (Table 3). In patients who received PN IRs of CRI were 34.84%; 38.89%, and 23.64% in the gastric, pancreatic, and lung cancer cohorts, respectively. While in patients who did not receive PN, IRs of CRI were 29.69%, 28.58%, and 22.13% in the gastric, pancreatic, and lung cancer cohorts, respectively (Table 4).
Table 3.
Incidence rate ratios of catheter-related infection (parenteral nutrition vs no parenteral nutrition) in cancer patients.
| Gastric cancer N = 5300 |
Pancreatic cancer N = 5975 |
Lung cancer N = 15,102 |
||||
|---|---|---|---|---|---|---|
| IRR (95% CI) | P-value | IRR (95% CI) | P-value | IRR (95% CI) | P-value | |
| Crude | 1.2473 (1.1330–1.3732) | <.0001 | 1.2404 (1.1304–1.3612) | <.0001 | 1.3211 (1.2093–1.4432) | <.0001 |
| Adjusted for all variables | 1.1165 (1.0119–1.2319) | .0281 | 1.2071 (1.0985–1.3265) | <.0001 | 1.1850 (1.0832–1.2962) | .0002 |
| Adjusted for age | 1.2423 (1.1283–1.3679) | <.0001 | 1.2438 (1.1330–1.3655) | <.0001 | 1.3279 (1.2154–1.4507) | <.0001 |
| Adjusted for sex | 1.2419 (1.1280–1.3674) | <.0001 | 1.2275 (1.1185–1.3471) | <.0001 | 1.3064 (1.1956–1.4274) | <.0001 |
| Adjusted for CCI (0/1/2/3–4/5+) | 1.2424 (1.1283–1.3680) | <.0001 | 1.2406 (1.1306–1.3615) | <.0001 | 1.2820 (1.1731–1.4011) | <.0001 |
| Adjusted for metastasis | 1.2184 (1.1063–1.3420) | <.0001 | 1.2541 (1.1426–1.3764) | <.0001 | 1.3203 (1.2086–1.4424) | <.0001 |
| Adjusted for history of infection | 1.1787 (1.0698–1.2987) | .0009 | 1.2090 (1.1015–1.3270) | <.0001 | 1.1905 (1.0884–1.3023) | .0001 |
| Adjusted for chemotherapy | 1.2013 (1.0900–1.3240) | .0002 | 1.2357 (1.1253–1.3569) | <.0001 | 1.3360 (1.2230–1.4595) | <.0001 |
CI = confidence interval, CCI = Charlson Comorbidity Index, IRR = Incidence rate ratio.
Table 4.
Incidence of catheter-related infection.
| Exposure group, n (%) | Gastric cancer N = 5300 |
Pancreatic cancer N = 5975 |
Lung cancer N = 15,102 |
|||
|---|---|---|---|---|---|---|
| PN = 2144 (40.45%) | No PN = 3156 (59.55%) | PN = 1836 (30.73%) | No PN = 4139 (69.27%) | PN = 2525 (16.72%) | No PN = 12,577 (83.28%) | |
| Incidence of CRI (%) | 747 (34.84%) | 937 (29.69%) | 714 (38.89%) | 1183 (28.58%) | 597 (23.64%) | 2783 (22.13%) |
| Sum of person-time of follow-up (in days) | 1,263,984 | 1,977,585 | 983,476 | 2,021,279 | 998,094 | 6,146,810 |
| Incidence rate of CRI per 1000 patient-days (95% CI) | 0.5910 (0.5501 -0.6349) | 0.4738 (0.4444–0.5051) | 0.7260 (0.6747–0.7812) | 0.5853 (0.5529–0.6196) | 0.5981 (0.5520–0.6481) | 0.4528 (0.4362–0.4699) |
| Sum of person-time of follow-up (in months) | 41527.2 | 64972.0 | 32311.3 | 66407.5 | 32791.6 | 201948.6 |
| Incidence rate of CRI per patient-month (95% CI) | 0.01799 (0.01674–0.01933) | 0.01442 (0.01353–0.01538) | 0.02210 (0.02053–0.02378) | 0.01781 (0.01683–0.01886) | 0.01821 (0.01680–0.01973) | 0.01378 (0.01328–0.01430) |
| Sum of person-time of follow-up (in years) | 3460.6 | 5414.3 | 2692.6 | 5534.0 | 2732.6 | 16829.0 |
| Incidence rate of CRI per patient-year (95% CI) | 0.2159 (0.2009–0.2319) | 0.1731 (0.1623–0.1845) | 0.2652 (0.2464–0.2854) | 0.2138 (0.2019–0.2263) | 0.2185 (0.2016–0.2367) | 0.1654 (0.1593–0.1716) |
CRI = catheter-related infection, CI = confidence interval, PN = parenteral nutrition.
3.4. Type of CRIs
The distribution of CRIs according to the adopted definition is presented in all 3 cancer types (Table 5) and by PN status (Table S1, Supplemental Digital Content, https://links.lww.com/MD/P127). The most frequently reported infections were those due to Escherichia coli and Coagulase negative staphylococcus. Coagulase negative staphylococcus was more prevalent in PN-exposed patients than in unexposed ones. Systemic inflammatory response syndrome was often reported in lung cancer (14.7%). Sepsis, especially, sepsis due to gram-negative bacteria was more frequent in pancreatic cancer (20.7%). Infections and inflammatory reactions from other cardiac and vascular devices and from other internal prosthetic devices, implants and grafts were more common in patients with PN.
Table 5.
Distribution of catheter-related infections by type.
| Type of infection, n (%) | Gastric cancer patients with infection N = 1,684 |
Pancreatic cancer patients with infection N = 1,897 | Lung cancer patients with infection N = 3,380 |
|---|---|---|---|
| Infections associated with medical device | |||
| Infection following a procedure, not elsewhere classified | 98 (5.82%) | 72 (3.80%) | 125 (3.70%) |
| Infection and inflammatory reaction due to other cardiac and vascular devices, implants, and grafts | 64 (3.80%) | 72 (3.80%) | 123 (3.64%) |
| Infection and inflammatory reaction due to other internal prosthetic devices, implants, and grafts | 30 (1.78%) | 35 (1.85%) | 41 (1.21%) |
| Infections following infusion, transfusion, and therapeutic injection | 3 (0.18%) | 6 (0.32%) | 10 (0.30%) |
| Shock during or resulting from a procedure, not elsewhere classified | 6 (0.36%) | 10 (0.53%) | 8 (0.24%) |
| Microorganisms responsible for bacteria associated with a vascular device | |||
| Escherichia coli | 300 (17.81%) | 316 (16.66%) | 467 (13.82%) |
| Coagulase negative staphylococcus | 167 (9.92%) | 146 (7.70%) | 258 (7.63%) |
| Staphylococcus aureus | 122 (7.24%) | 88 (4.64%) | 391 (11.57%) |
| Klebsiella pneumoniae | 89 (5.29%) | 106 (5.59%) | 173 (5.12%) |
| Enterococcus | 84 (4.99%) | 108 (5.69%) | 120 (3.55%) |
| Pseudomonas aeruginosa | 70 (4.16%) | 51 (2.69%) | 293 (8.67%) |
| Other Enterobacteriaceae | 15 (0.89%) | 15 (0.79%) | 53 (1.57%) |
| Sepsis and other catheter-related infection | |||
| Gram-negative bacteria sepsis | 197 (11.70%) | 392 (20.66%) | 293 (8.67%) |
| Non-specified sepsis | 194 (11.52%) | 251 (13.23%) | 464 (13.73%) |
| SIRS | 201 (11.94%) | 209 (11.02%) | 497 (14.70%) |
| Fungal sepsis | 23 (1.37%) | 18 (0.95%) | 33 (0.98%) |
| Sepsis due to unspecified staphylococcus | 20 (1.19%) | 19 (1.00%) | 30 (0.89%) |
| Contaminated substances | 1 (0.06%) | 0 (0.00%) | 1 (0.03%) |
SIRS = systemic inflammatory response syndrome.
4. Discussion
This population-based study analyzed the incidence of CRIs occurring in a real-world setting using a single and large population sample from the French health administrative database. The current study included only adult patients with gastric, pancreatic, or lung cancer.
This study highlights several risk factors independently associated with CRI on a population-based analysis: CCI score ≥ 1, PN administration, CT, presence of metastases, and history of infection before catheter placement. Male sex was associated with CRI in both pancreatic and lung cancers. Whereas age older 50 years was only demonstrated as a risk factor in lung cancer. In this study, PN was one of the independent risk factors for the incidence of CRI in adult patients with gastric, pancreatic, or lung cancer. PN was associated with an increased risk of CRI of 11.7% in gastric, 20.7% in pancreatic, and 18.5% in lung cancer. This is not a surprising finding given that PN is a growth factor for bacteria and fungi because of its components and especially lipid emulsions.[17,18] Nevertheless, recent studies have questioned this statement, since there is no evidence that, for patients in PN, eliminating the lipidic component would prevent this kind of complication.[19,20]
Moreover, the magnitude of this increase in risk has been much lower than previously reported. According to the results of 2 French studies conducted on patients with cancer, PN was associated with increase in CRI risk with hazard ratios of 6 and 28.5.[21–24] However, these 2 studies included only a small sample size (315 and 371 patients, respectively) and were methodologically very different from the current work.
On the other hand, we assessed the risk of CRI in patients on CT with PN versus CT without PN. After adjustment for confounding variables, the addition of PN to CT did not have an impact on CRI occurrence in the gastric cancer cohort, whereas there was 8% and 35% increase in the risk of CRI in pancreatic and lung cancer cohorts, respectively. The effect of PN on CRI was previously studied in digestive cancer patients who underwent CT in France.[6] In that study, patients on PN had a higher likelihood of CRI than patients without PN (Hazard ratio, 5.66; 95% CI: 2.86–11.23; P < .001).
In our study, the incidence and risk of CRI varied by cancer type, consistent with what has been reported in previous studies.[21,25–27] The highest incidence of CRI occurred in patients with pancreatic cancer. Indeed, pancreatic cancer was found as an independent risk factor for CVC-related infection as described previously.[21,27] Incidences of CRI found in our study ranged between 0.59 and 0.73/1000 days, depending on the site of the disease. However, caution must be exercised when comparing our results with the incidence reported in other studies. First, because of the numerous definitions used for CRIs in literature.[28] In France, CVC-related infections are defined according to the criteria of the expert consensus-based clinical practice guidelines.[29] Yet, in this study, it was not possible to validate the adopted ICD-10-based definition of CRI from the claims data and to distinguish true CVC-related infections from other possible sources of infections. Moreover, most of the studies on that topic performed in France reported higher incidences (2.8/1000 days and 1.34/1000 catheter days) but were conducted more than 15 years ago.[22,30] Whereas, more recent publications reported results similar to ours (0.76/1000 and 0.97/1000 catheter days).[31,32] Obling et al[23] reported, 11% had several infections, while 75% did not have any.
The most serious complications related to the infection are catheter-related sepsis. Although serious and sometimes fatal, catheter-related sepsis is rare, even in cancer patients receiving active oncology treatments. The results of a study published recently reported 0.29/1000 catheter days, with 8/761 treated patients requiring hospitalization and 1 death.[24] It should be highlighted that it is difficult to compare this study with the current evidence since we studied the CRI incidence/time rather than incidence per catheter days (considering catheter removal). According to our results, the incidence of CRI was lower than that reported previously; therefore, although important, the risk of CRI should not determine the administration of PN in these patients.
Our study showed a fair frequency of PN use in cancer patients, especially in gastric and pancreatic cancers, (40.5% and 30.7%, respectively). In contrast, PN use was lower in patients with lung cancer (16.7%). These figures were slightly different from those reported in a previous study of 154 hospital wards, where PN prevalence was 19.6% in gastric cancer, 24.3% in pancreatic cancer, and 8.1% in lung cancer.[33] These discrepancies may be due to the difference in the characteristics of the patients included in the 2 studies. While age, gender, and proportion of patients undergoing CT do not really differ, we observe a difference in tumor status, where the frequency of metastasis was higher in the present study, which may have had an impact on the use of PN usually recommended in advanced stage.[34–38]
In addition, our results revealed a high frequency of infections with Coagulase negative staphylococcus and Staphylococcus aureus. Indeed, these 2 microorganisms are usually the causative pathogens of CVC infections.[39] This result was also compatible with other studies.[2,6,22,30] It is important to note that the rates of E coli infections presented in our study (Table 5) are likely to be overestimated. In clinical practice these gram-negative bacteria are more commonly found as microorganisms causing infections in other sites mostly in neutropenic patients[40] and is not classically associated with CVC device. Furthermore, a study in lung cancer patients found only 8% of E coli infections that were attributed to catheter device.[41]
To our knowledge, this is the first study evaluating the relationship between CRI incidence and PN use in oncology performed by using individual and comprehensive national health insurance data covering the entire French population, with a large sample size. In support of this, the number of incident cancer cases found in this study was similar to those reported by The French National Cancer Institute in 2018,[42] with an estimate of 6600 new cases of stomach cancer, 14,100 of pancreatic cancer, and 46,300 of lung cancer, compared with an estimate of 6400, 10,000, and 33,500 stomach, pancreatic, and lung cancer patients identified per year in this study. The small discrepancy in numbers may be due to the inclusion and/or exclusion criteria applied in this study.
This study has some limitations that are inherent to all claims databases, as claims are collected primarily for the purpose of government reimbursement of health care. In this study some potential confounders of CRI encountered in cancer patients were not considered due to the nature of the database, such as the type of catheter device used or the site of the catheter insertion, or the use of sterilization methods, as described elsewhere.[30,43–45] In addition, clinical data such as World Health Organization performance status, weight loss, precise disease stage (rather than metastasis), or presence of neutropenia are not available in SNDS. There is also a potential for misclassification regarding patient ascertainment to either PN or non-PN group. Data related to PN in the inpatient setting is not exhaustive, as medical acts and medical devices are not always well coded in the database. Patients classified in the group of non-PN, could have received a PN that was not recorded in the database. In addition, CT may have been underestimated because codes for the administration of systemic therapy were not used, but only ICD-10 codes and disease-related group codes for CT sessions. Diagnosis codes used are found in the National hospital discharge database component of the SNDS (inpatient setting), therefore infections or diagnoses from the outpatient setting of care could not be included during the cohort identification and analysis. Finally, while coding errors or inaccurate disease classification may have occurred, SNDS diagnosis codes are widely used in France for patient-recorded information, so it is unlikely that such errors are frequent.
Most cancer patients in France receive PN at home.[46] Recent systematic review and meta-analysis of 22 cohort studies, 5 case-control studies, and 1 randomized clinical trial reported that adding an antimicrobial lock solution was beneficial for successful catheter salvage in home PN-dependent patients with a CRI. Successful salvage rates were different depending on the germ involved and highest for coagulase negative staphylococci, followed by gram-negative rods and S aureus.[47] Another systematic review of 1107 studies screened, and included data comprised of 45,695 catheter days showed that CRI levels were significantly decreased in patients using taurolidine compared to saline or heparin. The cumulative proportion of patients without CRI after 1 year was in 88% in Taurolidine group, 56% in Saline group, and 14% in Heparin group.[48] The European Society for Clinical Nutrition and Metabolism Guidelines on home PN recommend using strict aseptic technique for the care of home CVC device and use taurolidine locking as an additional strategy to prevent CRIs for its favorable safety and costs profile.[49] Furthermore, considering the alarmingly growing global threat of antimicrobial resistance and deaths, it is even more important to combine rigorous infection prevention with minimization of inappropriate use of antibiotics especially in cancer patients receiving CT and PN interventions.[50]
In conclusion, the results confirmed that PN was one of the independent risk factors for CRI. Nevertheless, the results of the current study suggested that the risk of CRI in patients with gastric, pancreatic, or lung cancer on PN was lower than reported previously, regardless of the patients were receiving or not adjuvant CT. However, considering that the data were extracted from a claims database, caution should be exercised when interpreting these results. With the intention of detecting and resolving potential complications, cancer patients receiving PN require close and personal monitoring for assessing treatment efficacy and safety. Therefore, to ensure the best possible quality of care, monitoring should be performed by healthcare professionals who have the necessary skills and tools, and use specific safety care protocols to evaluate, prevent and manage CRI across entire cancer patient care continuum (from hospital to homecare).
Acknowledgments
Ciencia y Deporte S.L. (A Coruña, Spain) has provided medical writing and editorial assistance.
Author contributions
Conceptualization: Julien Taieb, Olivier Pellerin, Gérard de Pouvourville, Jacques Massol, Laurène Gautier, Patricia Medina.
Formal analysis: Laurène Gautier, Patricia Medina.
Methodology: Julien Taieb, Olivier Pellerin, Gérard de Pouvourville, Jacques Massol, Laurène Gautier, Patricia Medina.
Writing – original draft: Julien Taieb, Olivier Pellerin, Gérard de Pouvourville, Jacques Massol, Séverine Jaskulski, Julian Shepelev, Laurène Gautier, Patricia Medina.
Writing – review & editing: Julien Taieb, Olivier Pellerin, Gérard de Pouvourville, Jacques Massol, Séverine Jaskulski, Julian Shepelev, Laurène Gautier, Patricia Medina.
Supplementary Material
Abbreviations:
- CCI
- Charlson Comorbidity Index
- CI
- confidence interval
- CRI
- catheter-related infection
- CT
- chemotherapy
- CVCs
- central venous catheters
- GI
- gastrointestinal
- ICD-10
- International Classification of Diseases, 10th revision
- IR
- incidence rates
- IRR
- incidence rate ratio
- PN
- parenteral nutrition
- SNDS
- French National Health Data System
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article. The authors, LG and PM disclose sponsorship from Baxter Healthcare SA, Switzerland for the research and publication of this article. The funding agreement ensured the researchers and authors’ independence in designing the study, interpreting the data, and writing and publishing the results.
At no point did the authors have access to patient-level data. Authors were provided only data and does not contain any identifiable information about real patients, so there is no risk of breaching patient confidentiality. Authors were provided aggregate results; post running of the analysis and is classed as secondary users to use and access to de-identified data compliant with the French data protection authority requirements.
The authors have no conflicts of interest to disclose.
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
Supplemental Digital Content is available for this article.
How to cite this article: Taieb J, Pellerin O, de Pouvourville G, Massol J, Jaskulski S, Shepelev J, Gautier L, Medina P. Incidence of catheter-related infection in cancer patients receiving parenteral nutrition: A retrospective cohort study of French administrative claims data. Medicine 2025;104:23(e42704).
Contributor Information
Olivier Pellerin, Email: olivier.pellerin@aphp.fr.
Gérard de Pouvourville, Email: pouvourville@essec.edu.
Jacques Massol, Email: jacques.massol.ext@aixial.com.
Séverine Jaskulski, Email: severine_jaskulski@baxter.com.
Julian Shepelev, Email: julian_shepelev@baxter.com.
Laurène Gautier, Email: laurene.gautier@hotmail.fr.
Patricia Medina, Email: patricia.m.medina@oracle.com.
References
- [1].Yilmaz G, Koksal I, Aydin K, Caylan R, Sucu N, Aksoy F. Risk factors of catheter-related bloodstream infections in parenteral nutrition catheterization. JPEN J Parenter Enteral Nutr. 2007;31:284–7. [DOI] [PubMed] [Google Scholar]
- [2].Beghetto MG, Victorino J, Teixeira L, de Azevedo MJ. Parenteral nutrition as a risk factor for central venous catheter-related infection. JPEN J Parenter Enteral Nutr. 2005;29:367–73. [DOI] [PubMed] [Google Scholar]
- [3].McCleary EJ, Tajchman S. Parenteral nutrition and infection risk in the intensive care unit: a practical guide for the bedside clinician. Nutr Clin Pract. 2016;31:476–89. [DOI] [PubMed] [Google Scholar]
- [4].Franck AJ. Quality improvement initiative to reduce adverse effects associated with parenteral nutrition overfeeding. Frontline Gastroenterol. 2019;10:67–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [5].Lafuente Cabrero E, Terradas Robledo R, Civit Cuñado A, et al. Risk factors of catheter- associated bloodstream infection: systematic review and meta-analysis. PLoS One. 2023;18:e0282290. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Touré A, Chambrier C, Vanhems P, Lombard-Bohas C, Souquet JC, Ecochard R. Propensity score analysis confirms the independent effect of parenteral nutrition on the risk of central venous catheter-related bloodstream infection in oncological patients. Clin Nutr. 2013;32:1050–4. [DOI] [PubMed] [Google Scholar]
- [7].van der Mee-Marquet N, Florent Goube F, Gimenes R, Valentin AS. Surveillance des infections liées aux dispositifs invasifs: Evolutions 2019-2022. Surveillance et Prévention des Infections Associées aux Dispositifs Invasifs (SPIADI). https://www.spiadi.fr/results?tab=0. Accessed December 3, 2023. [Google Scholar]
- [8].Cuschieri S. The STROBE guidelines. Saudi J Anaesth. 2019;13(Suppl 1):S31–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Bezin J, Duong M, Lassalle R, et al. The national healthcare system claims databases in France, SNIIRAM and EGB: powerful tools for pharmacoepidemiology. Pharmacoepidemiol Drug Saf. 2017;26:954–62. [DOI] [PubMed] [Google Scholar]
- [10].Tuppin P, de Roquefeuil L, Weill A, Ricordeau P, Merlière Y. French national health insurance information system and the permanent beneficiaries sample. Rev Epidemiol Sante Publique. 2010;58:286–90. [DOI] [PubMed] [Google Scholar]
- [11].ICD-10 Version:2016. ICD-10 Version:2016 (who.int). Accessed December 3, 2023.
- [12].Bousquet C, Trombert B, Souvignet J, Sadou E, Rodrigues JM. Evaluation of the CCAM hierarchy and semi structured code for retrieving relevant procedures in a hospital case mix database. AMIA Annu Symp Proc. 2010;2010:61–5. [PMC free article] [PubMed] [Google Scholar]
- [13].l’Assurance Maladie. Base des médicaments et informations tarifaires. http://www.codage.ext.cnamts.fr/codif/bdm_it/index.php. Accessed December 3, 2023. [Google Scholar]
- [14].l’Assurance Maladie. La liste des produits et prestations – LPP [updated March 1, 2023]. http://www.codage.ext.cnamts.fr/codif/tips/index_presentation.php?p_site=AMELI. Accessed December 3, 2023. [Google Scholar]
- [15].Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–83. [DOI] [PubMed] [Google Scholar]
- [16].Bannay A, Chaignot C, Blotière PO, et al. The best use of the Charlson comorbidity index with electronic health care database to predict mortality. Med Care. 2016;54:188–94. [DOI] [PubMed] [Google Scholar]
- [17].Opilla M. Epidemiology of bloodstream infection associated with parenteral nutrition. Am J Infect Control. 2008;36:S173.e5–8. [DOI] [PubMed] [Google Scholar]
- [18].Reimund JM, Rahmi G, Escalin G, et al. Efficacy and safety of an olive oil-based intravenous fat emulsion in adult patients on home parenteral nutrition. Aliment Pharmacol Ther. 2005;21:445–54. [DOI] [PubMed] [Google Scholar]
- [19].Gavin NC, Larsen E, Runnegar N, et al. Association between parenteral nutrition-containing intravenous lipid emulsion and bloodstream infections in patients with single-lumen central venous access: a secondary analysis of a randomized trial. JPEN J Parenter Enteral Nutr. 2023;47:783–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Tota A, Serra A, Raoul P, Gasbarrini A, Rinninella E, Mele MC. Lipid-Enriched parenteral nutrition and bloodstream infections in hospitalized patients: is it a real concern? Medicina (Kaunas). 2022;58:885. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Touré A, Vanhems P, Lombard-Bohas C, et al. Totally implantable central venous access port infections in patients with digestive cancer: incidence and risk factors. Am J Infect Control. 2012;40:935–9. [DOI] [PubMed] [Google Scholar]
- [22].Penel N, Neu JC, Clisant S, Hoppe H, Devos P, Yazdanpanah Y. Risk factors for early catheter-related infections in cancer patients. Cancer. 2007;110:1586–92. [DOI] [PubMed] [Google Scholar]
- [23].Obling SR, Wilson BV, Kjeldsen J. Home parenteral support in patients with incurable cancer. Patient characteristics of importance for catheter related complications and overall survival. Clin Nutr ESPEN. 2018;28:88–95. [DOI] [PubMed] [Google Scholar]
- [24].Cotogni P, Mussa B, Degiorgis C, De Francesco A, Pittiruti M. Comparative complication rates of 854 central venous access devices for home parenteral nutrition in cancer patients: a prospective study of over 169,000 catheter-days. JPEN J Parenter Enteral Nutr. 2021;45:768–76. [DOI] [PubMed] [Google Scholar]
- [25].Beckers MM, Ruven HJ, Seldenrijk CA, Prins MH, Biesma DH. Risk of thrombosis and infections of central venous catheters and totally implanted access ports in patients treated for cancer. Thromb Res. 2010;125:318–21. [DOI] [PubMed] [Google Scholar]
- [26].Napalkov P, Felici DM, Chu LK, Jacobs JR, Begelman SM. Incidence of catheter-related complications in patients with central venous or hemodialysis catheters: a health care claims database analysis. BMC Cardiovasc Disord. 2013;13:86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Lipitz-Snyderman A, Sepkowitz KA, Elkin EB, et al. Long-term central venous catheter use and risk of infection in older adults with cancer. J Clin Oncol. 2014;32:2351–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Tomlinson D, Mermel LA, Ethier MC, Matlow A, Gillmeister B, Sung L. Defining bloodstream infections related to central venous catheters in patients with cancer: a systematic review. Clin Infect Dis. 2011;53:697–710. [DOI] [PubMed] [Google Scholar]
- [29].Timsit JF, Baleine J, Bernard L, et al. Expert consensus-based clinical practice guidelines management of intravascular catheters in the intensive care unit. Ann Intensive Care. 2020;10:118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].L’Hériteau F, Olivier M, Maugat S, et al. Impact of a five-year surveillance of central venous catheter infections in the REACAT intensive care unit network in France. J Hosp Infect. 2007;66:123–9. [DOI] [PubMed] [Google Scholar]
- [31].Willsmore JD, Vanniasegaram D, Khanna P, et al. Central venous catheter salvage following a catheter related bloodstream infection in patients on home parenteral nutrition. Clin Nutr ESPEN. 2020;40:678. [Google Scholar]
- [32].Saqui O, Fernandes G, Allard J. Central venous catheter infection in Canadian home parenteral nutrition patients: a 5-year multicenter retrospective study. Br J Nurs. 2020;29:S34–42. [DOI] [PubMed] [Google Scholar]
- [33].Hébuterne X, Lemarié E, Michallet M, de Montreuil CB, Schneider SM, Goldwasser F. Prevalence of malnutrition and current use of nutrition support in patients with cancer. JPEN J Parenter Enteral Nutr. 2014;38:196–204. [DOI] [PubMed] [Google Scholar]
- [34].Rosania R, Chiapponi C, Malfertheiner P, Venerito M. Nutrition in patients with gastric cancer: an update. Gastrointest Tumors. 2016;2:178–87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [35].Arends J, Bachmann P, Baracos V, et al. ESPEN guidelines on nutrition in cancer patients. Clin Nutr. 2017;36:11–48. [DOI] [PubMed] [Google Scholar]
- [36].Kubota T, Shoda K, Konishi H, Okamoto K, Otsuji E. Nutrition update in gastric cancer surgery. Ann Gastroenterol Surg. 2020;4:360–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [37].Carrato A, Cerezo L, Feliu J, et al. Clinical nutrition as part of the treatment pathway of pancreatic cancer patients: an expert consensus. Clin Transl Oncol. 2022;24:112–26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [38].Kiss N, Curtis A. Current insights in nutrition assessment and intervention for malnutrition or muscle loss in people with lung cancer: a narrative review. Adv Nutr. 2022;13:2420–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [39].Mermel LA, Farr BM, Sherertz RJ, et al. Guidelines for the management of intravascular catheter-related infections. Clin Infect Dis. 2001;32:1249–72. [DOI] [PubMed] [Google Scholar]
- [40].Böll B, Schalk E, Buchheidt D, et al. Central venous catheter-related infections in hematology and oncology: 2020 updated guidelines on diagnosis, management, and prevention by the Infectious Diseases Working Party (AGIHO) of the German Society of Hematology and Medical Oncology (DGHO). Ann Hematol. 2021;100:239–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [41].Lanoix JP, Pluquet E, Lescure FX, et al. Bacterial infection profiles in lung cancer patients with febrile neutropenia. BMC Infect Dis. 2011;11:183. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [42].Les cancers en France en 2018 - L’essentiel des faits et chiffres (édition 2019. Updated February 5, 2019). https://www.cancer.fr/catalogue-des-publications/les-cancers-en-france-en-2018-l-essentiel-des-faits-et-chiffres-edition-2019. Accessed December 4, 2023. [Google Scholar]
- [43].Goetz AM, Wagener MM, Miller JM, Muder RR. Risk of infection due to central venous catheters: effect of site of placement and catheter type. Infect Control Hosp Epidemiol. 1998;19:842–5. [DOI] [PubMed] [Google Scholar]
- [44].Safdar N, Kluger DM, Maki DG. A review of risk factors for catheter-related bloodstream infection caused by percutaneously inserted, noncuffed central venous catheters: implications for preventive strategies. Medicine (Baltim). 2002;81:466–79. [DOI] [PubMed] [Google Scholar]
- [45].Parienti JJ, du Cheyron D, Timsit JF, et al. Meta-analysis of subclavian insertion and nontunneled central venous catheter-associated infection risk reduction in critically ill adults. Crit Care Med. 2012;40:1627–34. [DOI] [PubMed] [Google Scholar]
- [46].Danel Buhl N, Vaillant MF, Schneider S, Seguy D, Lescut D. Épidémiologie régionale de la nutrition artificielle à domicile chez l’adulte en France en 2019. Nutr Clin Métab. 2022;36(1, Supplement):S39. [Google Scholar]
- [47].Gompelman M, Paus C, Bond A, et al. Comparing success rates in central venous catheter salvage for catheter-related bloodstream infections in adult patients on home parenteral nutrition: a systematic review and meta-analysis. Am J Clin Nutr. 2021;114:1173–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [48].Wouters Y, Causevic E, Klek S, Groenewoud H, Wanten GJA. Use of catheter lock solutions in patients receiving home parenteral nutrition: a systematic review and individual-patient data meta-analysis. JPEN J Parenter Enteral Nutr. 2020;44:1198–209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [49].Pironi L, Boeykens K, Bozzetti F, et al. ESPEN guideline on home parenteral nutrition. Clin Nutr. 2020;39:1645–66. [DOI] [PubMed] [Google Scholar]
- [50].GBD 2021 Antimicrobial Resistance Collaborators. Global burden of bacterial antimicrobial resistance 1990-2021: a systematic analysis with forecasts to 2050. Lancet. 2024;404:1199–226. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.