D: Conclusion(s) of evidence (pediatric oncology patients)

* The letter refers to the specific GRADE assessment described in section C1.

** Frequent swimming defined as swimming more than once a month.

C: Results in pediatric and adult oncology patients 
C1. Evidence in pediatric oncology patients

In our literature search, one study was included for this clinical question. Robbins et al (1999) (12) showed that pediatric oncology patients who swim with central venous catheters (CVC) are at increased risk of catheter-related infections. Their population was described as pediatric oncology patients with tunneled central venous catheters. “Swimming” was defined as swimming at least once while the catheter was in place. 46 patients did not swim at all, 49 patients did go swimming. They all swam in a public or private pool, nine of them additionally went swimming in a lake.

There were 34 catheter-related infections in the swimmers group compared to 13 infections among the non-swimmers. However, when rates of infection were calculated as number of infections per catheter per month, there were no significant differences between the groups. The swimmers group had their CVC in place significantly longer than the non-swimmers group. They report 34 infections in a total of 843 months (34/843) in the swimmers group versus 13 infections in 506 months (13/506). This means rates of an infection of 0.04% in the swimmers group, and 0.025% in the non-swimmers group, with a risk ratio of 1.6 (95% CI not reported) which they did not consider statistically significant (significance calculated based on 95% CI, but confidence intervals are not reported) (12).

They reported separate results for tunnel or exit infections, defined as pus or erythema and tenderness on the exit site (exit infection) or throughout the whole tunneled portion of the catheter (tunnel infection). A bloodstream infection was defined as sudden onset of fever in association with a positive blood culture.
In the swimmers group, 20 tunnel or exit infections were reported in 843 months (rate of 0.02%), versus 8 infections in 506 months in the non-swimmers group (0.016%). Regarding blood stream infections (BSI), in the swimmers group 14 BSI’s were reported in 843 months (rate 0.016%) versus 5 BSI’s in 506 months in the non-swimmers group (rate 0.0095). These results both did not differ significantly (12).

Robbins et al (1999) also evaluated if patients who swam frequently were at higher risk of infection than patients who swam only once or twice all summer. “Frequent swimming” was defined as swimming at least once a month, which was compared to patients who swam infrequently or not at all. They reported 35 frequent swimmers, versus 60 infrequent and non-swimmers. Of these, 19 catheter-related infections were reported in the frequent swimmers group versus 27 among the infrequent and non-swimmers. Also for these groups, a significant imbalance was seen in months of catheter in place. Therefore, the rates were calculated per months of catheter in place. The total number of infections in the frequent swimmers group, was 19 per 579 months (0.03%), versus 27 infections per 770 months in the infrequent/non-swimmers group (0.03%). Tunnel or exit infections were seen in 6/579 in the swimmers group (0.01%) versus 16/770 in the infrequent/non-swimmers group (0.02%). Bloodstream infections were reported in 13/579 (0.02%) in the swimmers group versus 11/770 in the infrequent/non-swimmers group (0.14%). None of the results differed significantly (12).

Table 14: Study characteristics Robbins 1999

Table 15: Outcomes Robbins 1999 – Swimming

H: GRADE: Grade quality assessment swimming: design is case-controlled study, inconsistency not serious, indirectness not serious, imprecision serious (downgraded one level because of small study population), publication bias unlikely, downgraded 1 level because of serious risk of bias (selection bias unclear, attrition bias high, detection bias low, reporting bias unclear, confounding bias high, other bias high).

Table 16: Outcomes Robbins 1999 – Frequent swimming

I: GRADE: Grade quality assessment frequent swimming: design is case-controlled study, inconsistency not serious, indirectness not serious, imprecision serious (downgraded one level because of small study population), publication bias unlikely, downgraded 1 level because of serious risk of bias (selection bias unclear, attrition bias high, detection bias low, reporting bias unclear, confounding bias high, other bias high).

C2. Additional recommendations guidelines (adults)
From our search for existing relevant guidelines, no additional guidelines were included for this clinical question.

B: Clinical question, search and selection

B1: Clinical question:
What is the effect of social restrictions (regarding risk of infections) on infections (prevalence and infectious complications) and other outcomes in children with any type of cancer and/or after stem-cell transplantation?

P = Children (aged 0-18 years) with any type of cancer and/or after stem-cell transplantation receiving anti-cancer treatment with curative intent
I = Social restrictions regarding risk of infections (i.e. restriction in school attendance, kindergarten, visiting zoo or farm, pets, swimming (whirlpool, sauna visits), being in crowded places, public transport, intimacy, flowers, or other author defined social restrictions)
C = (No social restrictions)
O = Severe or mild infections (prevalence and infectious complications), quality of life, anti-cancer treatment-related complications (adjustments in therapy or delay), costs, mortality, event-free survival

B2: Literature search and study selection
For the complete methodology of this guideline development and the search strategy, we refer to pages 18-23.
Of 6038 unique citations identified in the literature search, one study with pediatric oncology patients (Robbins, 1999 (12)) was included. Evidence was extracted and assessed using the GRADE methodology.