Home parenteral nutrition (HPN) is indicated in patients who are unable to effectively use their gastrointestinal tract due to intestinal failure, motility disorders, obstructions, and so forth. HPN allows patients to leave the hospital and resume a near-normal lifestyle. However, HPN is technically and socially complex. Providing HPN demands a patient-centered care approach by a specialized nutrition support team that includes a nurse, dietitian, pharmacist, and physician.
The standard plan of care for safe and effective HPN therapy includes insertion of a central venous catheter (CVC) through which the custom-made parenteral nutrition (PN) prescription is infused with an electronic pump to accurately run PN overnight while the patient sleeps. The patient and family/caregiver must learn the mechanical intricacies of administering PN and the vigilant maintenance of the CVC prior to going home. Once home, patients are visited by community nurses to ensure care of the catheter and proper administration of HPN are maintained. Routine parameters in managing outpatient HPN include regular clinic visits, blood tests, and telephone triage if complications arise. Routine blood tests are done not only for the adjustment of the fluid, electrolytes, and nutrient admixture but also to monitor and prevent metabolic abnormalities such as hyperglycemia and kidney injury from dehydration as well as potential outcomes associated with long-term use, such as PN-associated liver disease. Among other complications, CVC infection is the most frequent and serious.
Overall, studies on HPN have reported the CVC infection rate to be between 0.44 and 4.2 per 1000 catheter days based on positive blood cultures or diagnosis with a mortality rate up to 9%.1,2 Scollapio et al2 reported a 10-year retrospective review of 225 HPN patients and found that 55% of rehospitalizations were due to HPN complications such as CVC sepsis. In that study, there was a rate of 1.13 to 4.2 infections per 1000 catheter days, and most infections were reported to be the result of catheter contamination rather than infusion of contaminated PN fluids.2 It was concluded that catheter-related bloodstream infection was one of the most significant and grave complications of HPN therapy.2 Another study conducted on hospitalized PN patients prospectively investigated 153 patients with CVC who were exposed or not to PN.3 Of these, 28 patients who had CVC-related infection were higher PN users compare to patients without infection (60.7% vs 34.4%; P=0.010).3 In that study, malnutrition, length of stay before acquiring a CVC infection, and sustained hyperglycemia were not significant risk factors, but the use of PN was found to be the only risk factor for CVC-related infections.3 Ireton-Jones and DeLegge1 reported a 5-year retrospective study of 72 HPN support program service centers with a database of an average of 4540 patients per year. The authors examined length of PN therapy, demographics, diagnosis, rehospitalizations, CVC infection, and mechanical complications.1 For CVC infection, the most common bacterium was coagulase negative Staphylococcus, with an infection rate of 0.44 to 0.84 per 1000 catheter days based on blood cultures.1 As their data were collected from various centers, the authors cautioned underreporting of line sepsis because not all were reported.1 Likewise, they were unable to identify risk factors associated with the CVC infections.1 On the other hand, Chang et al4 reported risk factors that influence the rate of CVC infection in 68 HPN patients from two centers (Vancouver and Toronto). Factors include infusion of medications and blood work collected from CVCs; patients with a higher number of dependents; patients receiving social assistance; being a part-time student; and the presence of a family member trained in HPN procedures.4 The difference in HPN programs' care system was not a factor.4 Another study from our group showed that PN was an independent risk factor for CVC infection, with an infection rate of 2.3 per 1000 catheter days, when a patient participated in a prospective cohort study of several home infusion groups for the Center of Disease Control and Prevention (CDC).5 This prospective cohort study used rigorous criteria and tight follow-up of patients and may have represented the true risk of CVC infection in HPN patients when compared to previous retrospective studies.5 Since that study, we implemented advanced patient education strategies about line management to reduce this rate.
We also established a web-based Canadian Home Parenteral Nutrition Registry to collect prospective data on HPN patients' demography, PN prescriptions, and outcomes such as survival, complications, and PN dependence, which led to several publications. One of these publications presented data on HPN-related liver and bone complications, but it was found that data regarding CVC infection was not well documented.6 Raman et al6 found that many patients live remotely and were hospitalized in their local institutions; thus it is difficult to get detailed information about admissions. They also found there was lack of in-depth data gathering. The current form asked only for incidence of line sepsis and occlusion at a given period.6 The form was not designed for information regarding symptoms, types of organisms, or overall management.6
The purpose of this study was to fill this gap by conducting a 5-year retrospective chart review in 3 HPN centers representative of the academic Canadian HPN programs focusing on CVC-related infections: the rate of infection and treatment management.
Methods
This observational study was based on a 5-year retrospective chart review of adult patients participating in the Canadian HPN Registry between 2005 and 2009 from 3 HPN programs in 3 provinces (Edmonton, Alberta; Toronto, Ontario; and Winnipeg, Manitoba). Patients who were lost to follow-up within that period were excluded. The centers had local research ethics board (REB) approval for the on-going collection and data entry into the web-based HPN registry, and all patients consented. A separate REB approval was acquired for the current study. Demographics such as age, gender, diagnosis, and length of time on HPN were collected from the registry. During chart review, data collection included symptoms, hospitalizations, type of organisms, investigations and results, treatment and management, and review of the patient/family's technique. Collected data were entered into a Microsoft Excel spreadsheet (Microsoft Corp, Redmond, WA). Data from the spreadsheet were imported into a statistical software program (SPSS Student Version 18; IBM, Armonk, NY) for further analysis. Descriptive statistics analysis using the Spearman correlation between length of time on HPN and number of infections was completed. To determine infection rate, the Health Canada formula was used (number of infections divided by number of device days multiplied by 1000).
Results
Demography
There were 81 adult subjects recruited. As shown in Table 1, the subjects' mean age was 52 years (65.6% female). Most patients had tunneled catheters (66.7%) and the second most preferred type was the peripherally inserted central catheter (PICC) (25.9%). Single lumen catheters (58%) were chosen over double lumen catheters (40.7%). Most CVCs were inserted in the radiology department (72.8%). The length of time a catheter was in place varied from a minimum of 14 days to a maximum of 4380 days, which provided a median length of time of 281 days (Q1:109 days, Q3: 638 days).
| Characteristic (n=81) | Mean±SEM/median (range) or frequency of population (%) |
|---|---|
| Age | 51.98 ± 1.71 |
| Gender | |
| Male | 28/81 (34.6%) |
| Female | 53/81 (65.6%) |
| CVC type | |
| Tunneled | 54/81 (66.7%) |
| PICC | 21/81 (25.9%) |
| Implanted | 6/81 (7.4%) |
| Number of lumen | |
| Single | 47/81 (58.0%) |
| Double | 33/81 (40.7%) |
| Unknown | 1/81 (1.2%) |
| CVC duration (days) | 281 (14, 4380) |
| CVC inserted in | |
| Radiology | 59/81 (72.8%) |
| Operating room | 7/81 (8.6%) |
| Unknown | 15/81 (18.5%) |
| Duration of HPN, months | 36 (1324) |
PICC=peripherally inserted central catheter
Figure 1 shows the primary indication for HPN is short bowel syndrome (SBS) (54.3%), followed by patients with other causes of malabsorption (fistula, celiac disease) (17.3%) or dysmotility such as scleroderma (16%).
Infection event
There were 144 infection events. As shown in Table 2, the majority of patients came to the emergency room (82/144) of their respective program center (77/144); others went to a different hospital (53/144) or the location of the point of care was not documented (41/144). The majority of patients with CVC infection had tunneled catheters (43.8%), followed by PICC (21.9%). For those requiring hospital admissions, the mean length of stay was 11 days.
| Mean±SEM/median (range) or frequency of population (%) | |
|---|---|
| Data source | |
| Center | 129/144 (89.6%) |
| Patient | 8/144 (5.6%) |
| Community | 7/144 (4.9%) |
| CVC type | |
| Tunneled | 78/144 (43.8%) |
| PICC | 39/144 (21.9%) |
| Implanted | 18/144 (10.1%) |
| Other | 1/144 (0.6%) |
| Unknown | 8/144 (4.5%) |
| Presentation | |
| Center | 77/144 (43.3%) |
| Another hospital | 53/144 (29.8%) |
| Other | 5/144 (2.8%) |
| Unknown | 9/144 (5.1%) |
| Department | |
| Emergency room | 82/144 (46.1%) |
| HPN clinic | 20/144 (11.2%) |
| Unknown | 41/144 (23.0%) |
| Other | 1/144 (0.6%) |
| Length of stay, days | 10.9 ± 2.3 |
In terms of symptoms, Figure 2 shows that fever (58.3%) and chills (38.2%) were the predominant clinical manifestations of CVC infection. Other frequently reported symptoms were feeling dizzy from hypotension, nausea, vomiting, and myalgia. Other symptoms (15.3%) were not specifically associated with CVC infection but were reported at the time of presentation, such as sore throat, cough, difficulty breathing, and burning on urination. Some patients presented with hypotension (9.7%), elevated white blood cell counts (9%), redness at the catheter exit site (10.4%), or with purulent discharge at the catheter exit site (9.7%).
Clinical Investigation
As shown in Figure 3, the most prevalent laboratory tests ordered were blood cultures (89.6%) from both a peripheral vein and the CVC in 71 of 144 cases of suspected CVC infection (49.3%). From the CVC, cultures were positive in only 31 of 144 cases (21.5%), and from the peripheral vein, only 6 out of the 144 suspected CVC infection cases were positive (4.2%). When removed, the central catheter tip tested positive for microorganism in 39 (27.1%) cases, while in 21 cases (14.7%) the catheter site swabs were positive for bacteria.
Blood, skin, and catheter tip microorganisms
The most freque nt microorganisms are presented in Figure 4 and were as follows: gram-positive bacteria: coagulase negative Staphylococcus (25.7%) and Staphylococcus aureus (5.6%); gramnegative bacteria: Klebsiella species (11.1%) and Escherichia coli (5.6%); and fungus (Candida sp) (7.6%). Other microorganisms found in the patients' blood were Enterococcus sp, Streptococcus sp, Enterobacter sp, and Pseudomonas aeruginosa. Figure 5 shows a microorganism that grew around the catheter exit site: gram-positive S aureus (3.5%) followed by coagulase negative Staphylococcus (2.8%). The catheter tip grew coagulase negative Staphylococcus (4.9%) and S aureus (3.9%). In some cases, the clinicians documented that despite the lack of blood culture test or negative results, there was a presumptive CVC infection (11.8%) based on patients' history of line sepsis (Figure 4).
Descriptive statistics analysis using the Spearman correlation between length of time on HPN, number of lumens, CVC type, and number of infections was completed.
As shown in Table 3, the length of time on HPN varied from a minimum of 1 month to a maximum of 324 months, which provided a median length of time of 36 months (Q1:10 months, Q3: 84 months). We found a positive correlation between the duration of HPN and the number of infections (r=0.423; P<0.001).
| Median (range) or N (%) | Median (range) or N (%) | Correlation coefficient, r | P-value | |
|---|---|---|---|---|
| Mean length of time on HPN (months) | 36 (1324) | 1.79 (0.29) | 0.423 | <0.001 |
| Number of lumens in patient | 0.159 | 0.156 | ||
| Single | 47/81 (58.0%) | 1.53 (0.35) | ||
| Double | 33/81 (40.7%) | 2.21 (0.49) | ||
| Unknown | 1/81 (1.2%) | 0 | ||
| CVC type in patient | 0.244 | 0.028 | ||
| Tunneled | 54/81 (66.7%) | 1.33 (0.29) | ||
| PICC | 21/81 (25.9%) | 2.90 (0.77) | ||
| Implanted | 6/81 (7.4%) | 2.00 (0.58) |
Similarly, we found a correlation between CVC type and number of infections (P<0.05), with a lower number of infections in those with a tunneled catheter. No statistically significant correlation was observed with number of lumens and number of infections in this study, but higher numbers of infections were observed in those with double lumen.
Clinical management
In more than 22% of CVC infection cases, combined antibiotics were prescribed or single antibiotic therapy such as cephalosporin (19.4%), vancomycin (17.4%), and ciprofloxacin (6.30%) was prescribed (Figure 6). Candida sp was treated with antifungal agents (5.6%). In other cases, data were not available.
An infectious disease (ID) specialist was consulted, which resulted in catheter removal. The new CVC that clinicians frequently chose was the tunneled type or PICC.
In Table 4, for those who had their CVC removed, there was a 7-day period without a CVC and with an average of 15 days before receiving a new catheter. While being treated, the patients did not have PN infusion for 7.5 days. When discharged from the hospital, patients had a follow-up blood culture 5 days after the last antibiotic dose to confirm infection clearance.
| Mean ± SEM/median (range) or frequency of population (%) | |
|---|---|
| Antibiotic treatment duration, days | 17.9 ± 1.2 |
| Number of days without CVC | 7.5 (2120) |
| Number of days prior to inserting permanent CVC | 15 (5150) |
| Number of days HPN held | 7.5 (1120) |
| Follow-up number of days | 5.13 ± 0.83 |
Discussion
PN offers lifesaving therapy. Depending on the primary indication, a patient can rely on HPN for many months, years, or even lifetime. The standard plan of care for safe and effective PN home administration includes the insertion of a long-term CVC intensive patient education, careful laboratory monitoring, and adequate follow-up care. Prior to going home, the patient and caregiver must learn the mechanical intricacies and vigilant use and maintenance of the CVC. Once home, there is an array of professional care required to ensure safe delivery of PN. Despite all of this and a reduction in infection rates compared to previous Canadian studies, our results show that CVC infection as an HPN complication remains prevalent and requires hospitalization, antibiotic treatment for more than 2 weeks, and frequent CVC changes, with interruption of PN infusion for about 7 days, all of which take significant health care resources.
According to Grant,7 PN patients have an increased risk for septic complications due to a number of factors, including mal-nutrition, frequent use of broad-spectrum antibiotics, and co-existence of infected wounds, urinary tract infections, and/or lung infections.7 Moreover, indwelling CVCs serve both as an entry point and as a site for microbial development, with the PN solution providing adequate growth medium.7 In addition, the patient and caregiver's lack of conscientious catheter care plays a role in CVC complications, especially sepsis.7
In the current study, tunneled CVC was the most frequently used catheter, but we did not see any difference in the rate of CVC infections between types of catheter. Based on a review by Sands,8 the most suitable catheter for HPN is the tunneled type because it offers the lowest rates of infection and thrombosis and it is easier to use. Similarly, in the recent European Society for Parenteral and Enteral Nutrition (ESPEN) guidelines in adults with chronic intestinal failure, ESPEN recommended that tunneled CVCs or totally implanted devices be used for long-term HPN.9
With CVC infection, patients sought medical attention at their local emergency department with typical clinical presentations of fever and chills. Other symptoms were dizziness from low blood pressure, headache, tachycardia, altered mental status, fatigue, nausea, and vomiting. These symptoms are recognized by Band10 to be identical to those with bloodstream infection from other sources. For this reason, further investigations to exclude other sources of infection, such as in the lung and urinary track, is required. Although not all patients with CVC infections were hospitalized in our study, in other studies, including Campo et al,11 patients were admitted as a standard of care when they presented with fever and no evidence of other sources of infection due to the potential fatal outcome of CVC infection.
In terms of clinical investigation, blood cultures from both a peripheral vein and the CVC were performed in most patients, but some clinicians tested blood only from either a peripheral vein or the CVC. In Band's10 review of clinicians' practice, the author observed that blood cultures were frequently collected only peripherally and not frequently from the central catheter.10 Band recommended taking blood samples from the CVC as the sole source of blood culture but stated that caution should be used to avoid skin contamination as this may contribute to false-positive blood cultures.10 This contradicts the CDC position and recommendation to compare sequential blood cultures obtained from both a peripheral vein and the CVC, which is helpful in diagnosing catheter-related bloodstream infection.12 The position was largely due to a recent project completed by Canadian hospitals in collaboration with the CDC. The CDC12 strongly recommend that blood culture specimens should be taken from separate venipuncture sites and not through a CVC site. Other recommendations include semiquantitative culture of the skin and the catheter tip as a valuable method for implicating the CVC as the source of bacteremia, especially if cultures from the exit site or pocket reveal the same organism isolated from peripheral blood cultures.12
Blood culture results in the current study were similar to previous studies, specifically for coagulase negative Staphylococcus as the most frequently seen microbe in the blood from the skin or CVC tip, followed by S aureus and Klebsiella pneumonia.1,7,13,14,15 However, in our study, despite similar types of bacteria and clinical presentation, management varies between physicians as observed in different centers.
The standard of care is to use intravenous (IV) antibiotics via the CVC.12 However, on occasion, HPN programs see patients initially assessed by general practitioners who are not familiar with CVC infection and who tend to use oral antibiotics with poor results. The type of antibiotic is also important. Singhal et al13 studied the species distribution and antimicrobial susceptibility of coagulase negative Staphylococcus.13 The authors concluded that all microbe isolates were sensitive to various types of antibiotics, including vancomycin, teicoplanin, and linezolid.13 Santarpia et al14 considered the advantage of combining multiple antibiotics as a more effective treatment when managing catheter-related bloodstream infection. The authors concluded that a combination of two or more antibiotics avoids bacterial resistance.14
Consultation with an ID specialist is also important, particularly if the patient does not respond to treatment, deteriorates, or there is consideration for CVC removal. This may also help concerning length of time on appropriate antibiotics and timing for reinsertion of a new CVC. In a review by Hall et al,15 CVC removal was shown to avoid CVC infection-related morbidity, especially when there was no patient improvement or the bacteria could not be eradicated. CVC removal has been also associated with quicker recovery.15 In one center, a PICC was inserted as a temporary intravenous access until a new, tunneled CVC was inserted after antibiotic course completion and bacterial eradication.15 ESPEN's most recent recommendation regarding catheter management is the conservative approach with systemic and local antibiotic coverage.9 Moreover, catheter removal should be the first action in case of tunnel infections or positive blood cultures.9 Catheter removal is required for port abscesses, complex infections, unrelenting hemodynamic fluctuation, and blood cultures positive for fungi.9 When selecting a new CVC, the American Society for Parenteral and Enteral Nutrition's (ASPEN's) recent guidelines declare tunneled CVCs to be suitable for adult patients requiring long-term PN.16 The presence of a cuff within the tunnel prevents microbial migration and lessens risk of dislodgement.17 Furthermore, tunneled CVCs do not restrict upper extremity movements, the position on chest facilitates self-care, and it can be easily hidden under clothes.17
Another important aspect is patient/caregiver education to emphasize the importance of adhering to aseptic technique. By explaining the potential risks of line sepsis and that strict adherence to protocols and close monitoring may help decrease sepsis rate, the patient/caregiver may come to understand that the result will help increase the catheter longevity and improve quality of life. Santarpia et al14 reported that patient/family education on sepsis prevention plays a pivotal role in the management of line sepsis and is fundamental to the management of catheter-related outcomes. The ASPEN's standards for specialized nutrition support specifically for HPN patients require that patient and caregiver's education program must include infection control and prevention, including aseptic technique during PN administration via CVC, scrupulous hand-washing techniques, and precautions to avert infection.17 The most recent ASPEN Safety Consensus Recommendations remind clinicians that the patient's and caregiver's skill and adherence must be periodically reviewed and documented.18 For patients who repeatedly present with CVC-related infections, ESPEN recommends retraining of the patient/caregiver and/or use of an antimicrobial catheter lock.9
Others also encourage the use of interactive and video-taped material, which has been shown to reduce the incidence of infection, reactive depression, and rehospitalization due to infection.19 Brock et al20 evaluated the effects of using a personal digital assistant (PDA) device for educating patients with HIV/AIDS. They found there was statistically significant improvements in knowledge of the disease, medications, and adherence behaviors after the subjects watched the PDA-based video.20 Although they are sophisticated, perhaps technology-enhanced teaching programs are not the most effective method to teach adult patients, as evaluated by Nicol et al.21 Conceivably, because the HPN training program is more complex compared to Brock's PDA video, there could be a role for a technology-enhanced learning program integrated with traditional teaching style. In a systematic review by Rowe et al,22 the authors acknowledged clinical education, akin to health care education of patients and families, is a multifaceted process. Moreover, even though clinical education through online interaction is effective, adult students also need theory of knowledge, methods, validity, scope, and the distinction between justified belief and opinion.22 The authors concluded that integrating technology-enhanced teaching with traditional approaches has the capacity to improve competencies among the adult learners.22
Conclusions
In this 5-year retrospective chart review of 81 HPN patients, the CVC infection rate was 0.97 per 1000 catheter days, which was lower than previously reported by a Canadian program, suggesting improvement in practice and better patient education on preventive measures.5 CVC infection remains a serious complication that frequently requires hospitalization, multidisciplinary care, several investigations, CVC removal and reinsertion, as well as interruption of PN, all using significant health care resources. Repeatedly, the tunneled CVC has been postulated to be the most appropriate long-term catheter in this unfortunate subset of patient population. Equally important is patient education, and emphasis on catheter care technique to reduce this serious complication through various methods of learning is of paramount importance. Encouraging and connecting patients and families with support groups will also enhance their knowledge, skill, and confidence to live as normal a life as possible.
Limitations
This study has limitations that include the retrospective and observational nature of the design as well as the limited sample size of patients from 3 HPN programs. Data were not collected for those patients who were lost to follow-up or died. This may have affected the infection rate and therefore introduced a bias. The infection rate may be underestimated if some of these patients had CVC infection. However, these academic programs have similar patient populations and standards of care based on previous publications from the Canadian HPN Registry. In addition, the number of HPN patients is small, as this falls in the “rare disease” category, and Canadians on HPN are few compared to other countries such as the United States. Another limitation is the limited availability of all clinical information regarding CVC infection events as HPN patients are distributed over a vast geographical area and often seek medical care from local community hospitals.
Recommendations for practice
The following are general recommendations for recognizing, treating, and preventing CVC infection.
Further research is needed in the different education strategies one can employ to avoid CVC infection. In the age of modern technology, written and verbal instructions are not adequate. Perhaps incorporating electronic learning tools into the teaching design and methods, blending them with traditional teaching, can assist in understanding how patients and families learn best to avert CVC infections in this unfortunate subset of patient population.