Effectiveness.

Study findings of the RCTs and prospective cohort studies support the recommendation that subcutaneous hydration is at least as effective as IV therapy and may be superior in certain circumstances for the treatment of dehydration in the pediatric and elderly population (Table 4 and S6 Table. Both biochemical (e.g., urea and creatinine) and clinical improvement (e.g., general wellbeing and improvement in mentation) were reported in all studies investigating this outcome, in both SC and IV groups. One small prospective cohort study reported superiority in the SC group in urea (p = 0.001), creatinine (p < 0.001) and sodium (p < 0.05), with an overall clinical improvement rate of 77%. The only study showing a preference for IV therapy in 1 outcome measuring volume infused was in the pediatric population, in which patients who received SC hydration in the emergency department had to be switched to IV upon admission to inpatient units due to lack of staff ability to perform SC hydration, which led to reports of “0 ml” infused [50]. The study did demonstrate a statistically significant higher volume of infusion with subcutaneous infusion in the emergency department.

In the elderly population, improvement was similar to IV hydration, with two systematic reviews providing evidence that SC hydration may be the treatment of choice for patients with cognitive impairment [43, 47, 48]. Rationale cited include the ability to place catheters in discreet sites out of sight, relatively pain-insensitive sites, (although studies site comparable discomfort), ease of placement and indirect savings from not having to observe the patient as closely [43].

In the pediatric study, there was a 20% placement failure rate of children randomized to IV therapy vs no SC access failures [50]. A meta-analysis supports this finding, reporting that children and adult patients (n = 238) receiving IV hydration were more likely to experience an insertion failure (RR 14.79, 95% CI 2.87 to 76.08; GRADE rating: moderate) [33]. A systematic review concluded that the odds of correct initial needle placement was 7.19 times (2.93, 17.68), p = .04) higher in hyaluronidase-facilitated subcutaneous access than IV [34].

Both hydration and medication reviews included the use of hyaluronidase, an enzyme that facilitates subcutaneous absorption [15, 23, 27, 28, 33, 34, 39, 40, 42, 48–50]. Spandorfer and colleagues [50] report that hyaluronidase is safe, effective and well tolerated in subcutaneous hydration solutions for pediatrics. They found that gravity-driven infusion rates can be increased four-fold with better systemic fluid absorption. In the palliative care population, 150 units of hyaluronidase was injected as a bolus into the infusion site immediately before starting each twice daily 500 mL bolus over one hour of two-thirds dextrose and 1/3 normal saline [50]. No significant differences were observed for pain, swelling, edema, rash or leakage between placebo and hyaluronidase, with the authors concluding that hyaluronidase is not necessary for routine boluses, unless the patient does not tolerate the infusion well due to swelling or pain. Administration protocol varied across studies. In one study 150–300 units/L were added to the solution for continuous infusions; however, for hydration boluses, 150 units was injected into the SC site before the infusion [51]. A protocol for pediatrics was to insert the SC set and then flush it with 140/150 units hyaluronidase prior to starting the infusion and repeating the dose once every 24 hours during infusion, as needed [27, 34]. Other protocols included using hyaluronidase only if the infusion was running slowly (method not reported) [52] or if problems with resorption as a bolus injection before starting the infusion [53].

Safety.

Evidence suggests that subcutaneous hydration in the pediatric and elderly population is safe and equivalent to IV hydration, in terms of local site reactions and systemic events and superior in catheter extraction and device-related agitation (see Table 4).

In the pediatric population, most patients in both SC (100%) and IV (90%) experienced at least one adverse event (no serious events) [33]. There were fewer case of erythema (RR 0.43, 95% CI 0.31 to 0.61) and edema at the insertion site (RR 0.42, 95% CI 0.25 to 0.72; n = 453; p = 0.001). The majority were local site reactions (swelling, erythema and pain), which Spandorfer et al. [45] reported were resolved spontaneously without treatment. These authors also reported the pain as being related to needle insertion in both groups with resolution post-insertion. Duems Noriega and Arino-Blasco [16] report that these local reactions are frequent but easily avoidable (through rate and volume control, proper needle placement [avoiding muscle] and aseptic technique). Ker et al.’s meta-analysis [33] found no evidence that the number of patients reporting pain differed between IV and SC (RR 1.01, 95% CI 0.83 to 1.22; n = 262; p = 0.94). More patients experienced phlebitis with an IV device than SC device (RR 5.04, 95% CI 1.14 to 22.30; n = 181). Infection (e.g. cellulitis and lymphangitis) rates were higher in patients with IVs than SCI (RR 3.70, 95% CI 1.06 to 12.88; n = 211; p = 0.04) [33].

The most common, although rare, systemic reactions reported to be possibly related to SC therapy include hyponatremia in a total of 3 patients, cardiac failure in 2 patients, and fluid overload (mean 0.003 episodes per day of therapy), all of which were comparable rates to IV [46, 44, 52, 53].

Catheter dislodgments, in which the catheter/needle is pulled out, were more likely to occur with IV than SC in the elderly population (RR 3.78, 95% CI 1.16 to 12.34, p = .03) [34]. Placement of the SC catheter in this population in discreet locations, such as the upper back (which is out of sight and less accessible) leads to less dislodgements [43].

Acceptability.

Subcutaneous hydration is significantly favoured over IV hydration with regards to successful and ease of device placement, patient agitation, and patient/caregiver satisfaction (Table 4). Patients with an IV are more likely to be agitated than with SC [33, 43].

In an RCT of 96 elderly patients, 17 patients were switched to SC (from IV) due to lack of venous access (n = 8), removal of cannulae by patients (n = 5), or other causes not reported [47, 53]. The need for medications and lack of medications suitable for SC access is an impediment, as 11 patients had to be switched to IV access to accommodate medication administration thus hydration was also converted. One study showed no nursing preference for SC vs IV; however, it is noted that the study protocol was for physician placement of IV catheters (with the latter reporting SCI as significantly more feasible [54].

Ker and colleagues [33] suggest that the ease of administration and cost-effectiveness as being particularly useful for healthcare workers with minimal medical training. In a multi-centre pediatric study, most participating centers did not have protocol or training for inpatient staff [50]. Therefore, upon transfer from the emergency department to the inpatient unit, patients had to be switched to IV therapy, which led to designation of SCI treatment failures. This finding was echoed by a qualitative study which found that 71% of nurses were unaware of this technique and 100% reported not having received any organizational guidance [39].

Efficiency.

Subcutaneous hydration is significant favoured over IV hydration in terms of cost effectiveness, in a few measured outcomes (Table 4). Total treatment time (from first catheter insertion attempt to end of infusion) was reported in a pediatric study as significantly shorter than the IV route [50]. Two studies reported cost savings with SCI route, due to fewer and less costly cannulae required, although these did not reach statistical significance [43, 52]. Indirect costs are compounded by the additional supervision required for IV infusions [46]. Cost of IV therapy has been reported to be four times higher than SC therapy [39, 46].

Effectiveness.

In general, medications administered subcutaneously are absorbed slower than given IV; however, can reach similar plasma levels [16]. Medications in the reviews were generally found to be equivalent in efficacy as when given via other routes such as intravenous or orally. Pain medications have the most extensive reported SCI use. Extensive evidence was reported for the use of subcutaneous morphine compared to intravenous and multiple other routes of administration demonstrating good effectiveness and safety (Table 5). Subcutaneous infusions were reported compared to epidural and intravenous of various opioids finding them to be equivalent or equivalent in efficacy, safety and acceptability [48].

For pain management, SCI of morphine was compared with epidural infusions and found to be acceptable and effective with fewer hypotension concerns with subcutaneous than epidural [59, 60]. Favorable subcutaneous use has been reported with baclofen, buprenorphine, hydromorphone, ketorolac, ketamine, methadone, morphine, tramadol [16, 48]. Both intravenous and epidural morphine patient-controlled analgesia (PCA) were evaluated in 5 studies and reported to be comparable [48].

Two antibiotics (ceftriaxone and ertapenem) were found to have similar antimicrobial effect, although variances did occur with Tmax and Cmax values and need to be taken into consideration if time to maximal plasma concentration and serum concentration are important clinically [16, 44]. For diuretics, overall clinical results were found to be adequate when compared to intravenous administration of the same medication, although onset of action may be delayed which can limit use in emergency situations [16, 48]. For endocrine medications, efficacy evaluated by return to expected circadian rhythms for cortisol was acceptable with less adverse events reported than other routes of administration [16, 24, 48]. Anti-epileptics were evaluated based on blood levels and seizure control and found to be adequate [16, 61]. Anti-nausea medications and neuroleptics were evaluated based on symptom improvement. Anti-nausea medications were found to show decreases in nausea and vomiting ranging from 62–93% depending on medication and patient population [16, 62]. One matched cohort trial found metoclopramide to be significantly less tolerated than ondansetron when administered by continuous subcutaneous infusion (4.4% versus 31.8%, p<0.001) [62].

Midazolam is considered the safest choice for SCI as it is the only water-soluble benzodiazepine and it has a fast onset, good tolerance and short half-life while still being reversible by antidote flumazenil even after administration [16]. Chemotherapy agents and biologics, such as Trastuzumab, administered by SCI are limited but those reviewed were found to have good bioavailability, similar terminal half-lives and found to be equivalent or non-inferior in response rates to intravenous routes [16, 38]. SCI of desferrioxamine has been shown to be effective with high levels of adherence and still considered a well-established method of iron chelation [15,25].

SC immunoglobulin (SCIG) infusion was found to achieve acceptable trough levels and reported effectiveness similar to IV infusion [29]. SCIG was also associated with higher health related quality of life and reduced infection rates which both are used to evaluate effectiveness of immunoglobulin therapy.

SCI of treprostinil for pulmonary hypertension was compared to placebo and found to be significantly more effective with minimal and tolerable side effects but no reviews included a comparison to intravenous [37]. The reviews showed statistically significant improvement in mean pulmonary artery pressure (mPAP), pulmonary vascular resistance (PVR), right arterial pressure (RAP) and cardiac index (CI). Changes in exercise capacity and dyspnoea were statistically significant but uncertain clinical significance. Quality of life scores reported higher in study group than placebo. While intravenous studies against placebo showed greater magnitude of clinical benefit versus the magnitude seen between subcutaneous and placebo, the intravenous studies had a 12–25% reporting of serious adverse events (e.g. sepsis, pulmonary embolism, hemorrhage) with only minor events reported in the subcutaneous group. While reported in both the placebo group and the treprostinil group, infusion site pain and redness was reported as statistically significant for treprostinil but bleeding and bruising was not reported to be statistically significant [37]. Adverse events related to the medication such as diarrhea, jaw pain, vasodilation, and edema were statistically significant compared to placebo but very similar among routes of administration. While mortality was not a reported outcome in the subcutaneous trials included in the reviews, the improvement in RAP and CI reported are considered reliable indicators of survival for pulmonary hypertension per the European Society of Cardiology [26].

Safety.

Most commonly reported adverse events were local tissue reactions including redness, pruritus, and itching at the administration site (Table 5). Overall, the reviews reported adequate to good tolerance for the administration of the medications reviewed, once side effects attributed to the medication via any administration route were eliminated. However, caution must be taken, and patients advised to report, any adverse events, as studies have been very limited for most medications. Specific infusion rates and diluents vary based on medication and patient characteristics and could potentially impact tolerance. For example, in the treprostinil study, a placebo of diluent only was used and some site reactions still reported by placebo group but analysis not provided as to whether this occurred at lower doses or only with titration [26, 37].

Medications contraindicated for subcutaneous infusion administration include severe irritants such as higher supplements of potassium chloride (contraindicated as bolus but slow rates may be tolerated), lipo-soluble drugs such as diazepam which can precipitate in the skin, and medications known to cause fat necrosis such as chlorpromazine and prochlorperazine [16].

In cases of high osmolarity medications, such as subcutaneous immunoglobulins (SCIG), the subcutaneous route was considered safer as the infusion had less cardiovascular risks associated with the infusion. Reactions were more limited with the SCIG infusion versus the intravenous route with local skin whelps and irritation being the most frequent complaint. Specifically, in one review, 15 studies (376 patients) reported comparing adverse events during treatment for patients with Primary Antibody Deficiencies (PAD) with SCIG and IVIG, respectively [29]. The calculated odds ratio of 0.09 (0.07–0.11; p<0.001) indicates a significant preference of SCIG over IVIG because of a decrease in systemic adverse events. The data analysis clearly demonstrates that SCIG therapy, independent of the rate of infusion (standard, “rapid” or “express” administration), is associated with a high incidence of local reactions (44.7% for SCIG 10%, 92% for SCIG 16%, and 100% for SCIG 20% vs. 32% of IVIG 10%).

Acceptability.

In one large systemic review, 5 of 94 studies reported preference for subcutaneous over IV with reasons varying from less side effects, to less time involved, less pain with access/administration or overall comfort [16]. SCIG patients also reported acceptability and a preference for subcutaneous. In one review specifically designed to evaluate patient acceptance of SCI versus intravenous medication administration, acceptability or preference for subcutaneous was found in 4 of 6 studies, ranging from 44–91% preferring SCI over intravenous with no significant difference found in one study and one showing a small preference for intravenous [38]. Reported reasons for the SCI preference included home/independent administration, convenience, less administration discomfort, and no end-of-dose weakening. Reasons of the group that preferred IV was not reported.

Efficiency.

Few studies specifically reported on resource savings. In the review of clinical efficacy for diuretics, hospitalization of those patients with poor venous access was avoided in 83–93% cases [16]. One report on SCI of anti-emetics found it to be cost prohibitive versus other therapies (not versus other routes of administration) but since the time the article was published, the medication cost has drastically changed [62].

Shifting from hospital based IVIG treatment to home based SCIG is reported as “25–33% of the annual costs (US$11,000 per patient per year) in Sweden, 50% of the annual costs (€17–77 million) in Germany, CA$2,000 per patient per year in Canada (total cost savings CA$ 9,000,000/y), 25–50% of the annual costs in France, and $2,000–5,000 per patient per year in the US” ([29], p.188).