Key points from the evidence

3. The most recent systematic review (53 studies; two comparative; n=2,218) compared patients receiving bacteriophage therapy with a control group receiving standard of care or placebo across several medical and surgical specialties.³ Based on a pooled analysis of randomised controlled trials (RCTs), patients in the control group were more likely to show clinical improvement than patients who received bacteriophage therapy (50% vs 43.8%; two RCTs). Conversely, patients treated with bacteriophage therapy were more likely to achieve bacterial eradication than those in the control group (16.6% vs 0.0%; 1 RCT).

4. A second systematic review (20 studies; 0 comparative; n=51) outlined the effect of bacteriophage therapy, with or without conventional antibiotic therapy, on patients with bone and joint infections.⁴ Criteria for success following treatment were satisfied in 71% of treatment episodes; success by indication was 57% for patients with periprosthetic joint infections and 88% for patients with osteomyelitis.

5. A third systematic review (27 studies; two comparative; n=1,579) reported results for the effectiveness of bacteriophage therapy, with or without conventional antibiotic therapy, in patients with three different infection types:⁵

a. burn wound infections: 49.6% of patients achieved clinical resolution, 27.9% showed improvement, and 22.5% showed no improvement. Excluding three studies that did not clearly report on clinical resolution, or where the bacteriophage therapy dropped below the therapeutic dose, resulted in 89.2% of patients achieving clinical resolution.

b. chronic wound or ulcer infections: 65.8% achieved clinical resolution, 20.3% showed improvement, and 13.9% showed no improvement.

c. dermatological infections: 87.3% achieved clinical resolution, 6.8% showed improvement, and 5.9% showed no improvement. 

6. The fourth systematic review (30 studies; three comparative; n>1,152, one study did not report patient numbers) reported results for bacteriophage therapy in patients with multidrug resistant infections.⁶ Twenty-six of the 30 studies included showed that bacteriophage therapy successfully decreased or halted bacterial growth.

7. Safety data on bacteriophage therapy came from the four systematic reviews described under the clinical effectiveness subheading above.^3-6 Not all of the studies included in the reviews reported safety outcomes, hence the number of included studies and patient numbers differ between the clinical effectiveness and safety sections.

8. The most recent systematic review (51 studies; nine comparative; n=731) found that, based on a pooled analysis of nine RCTs, patients treated with bacteriophage therapy were less likely to experience adverse events than those in the control group (7.6% vs 14.9%; 9 RCTs).³ No adverse events were reported in the observational studies. A pooled analysis of case series found an adverse event rate comparable to that observed in the RCTs.

9. The second systematic review (20 studies; n=51) noted that eight of the 20 studies included reported adverse events associated with bacteriophage therapy, finding that adverse events occurred in 8% of treatment episodes, all of which were considered to be minor: elevation of liver function tests, mild pruritis associated with an elevation of Tumour Necrosis Factor alpha, or redness and pain.⁴

10. The third systematic review (15 studies; n=1,095) included three studies in patients with burn wound infections that found no adverse events, five studies in patients with chronic wound or ulcer infections, four of which identified no adverse events, and eight studies in patients with dermatological infections, seven of which did note adverse events.⁵ Adverse events were described on an individual study basis, were relatively mild and were not thought to be directly related to bacteriophage therapy. Adverse event data from this review was primarily from studies conducted between 1929 and 1987 and therefore should be interpreted with caution.

11. The fourth systematic review (22 studies; n=not reported) reported that 20 studies found no adverse events.⁹ The remaining two studies noted subsequent infections after treatment, eczema, increased pain, nausea and vomiting, but had limited data to confirm whether these were related to bacteriophage therapy. No studies reported an association between phage administration and death.

12. The literature on patient and social aspects is limited to a single study on patient preferences around bacteriophage therapy and antibiotic therapy for diabetic foot ulcer infections. The study explored patient awareness and concern about antibiotic resistance, and perceptions of bacteriophage therapy, through a survey (n=55) and focus group (n=5) with Scottish patients and found that:

a. patients’ levels of concern about antibiotic versus bacteriophage therapy were not statistically significantly different

b. after reading information about bacteriophage therapy, 87% of patients stated they would accept bacteriophage therapy if it was recommended by their doctor

c. all focus group participants were supportive of bacteriophage therapy, and four of the five strongly expressed a willingness to use bacteriophage therapy in lieu of intravenous antibiotics if possible, citing ease of use, the potential not to be admitted to hospital and the likelihood of a significantly reduced side effect profile.

13. The views of patients were captured via a patient organisation submission received from Antibiotic Research UK which noted that:

a. patients with antibiotic resistant infections feel there is a lack of acceptance or confirmation of their ongoing suffering and a need to travel to private clinics in England, or even abroad, to access alternative treatments

b. recurrent and resistant infection's dramatically reduce quality of life for patients and their families. They severely affect mental health; many patients say life is not worth living and admit to having suicidal thoughts.

c. antibiotic resistant infections represent an economic burden for patients through being unable to work and to the NHS due to long-term care needs.

14. No published cost effectiveness evidence was identified during the literature search. An economic model was therefore developed by SHTG to inform the recommendations on the use of bacteriophage therapy for the treatment of patients with difficult to treat bacterial infections.

15. The economic evaluation estimated the cost effectiveness of bacteriophage therapy plus standard of care versus standard of care only for the treatment of adults with severe treatment-refractory diabetic foot infections who were at risk of amputation despite conventional antibiotic therapy. This population represents a subset of the overall patient population who may be suitable for bacteriophage therapy, and was selected for analysis based on the availability clinical expertise to inform the analysis.

16. Base case results indicate that bacteriophage therapy plus standard of care is less costly and more effective than standard of care only. The cost savings associated with bacteriophage therapy plus standard of care stem from a lower proportion of patients requiring minor or major lower extremity amputation, leading to an overall lower cost of care and higher quality-adjusted life years (QALYs). Of note:

a. probabilistic sensitivity analysis estimated that bacteriophage therapy plus standard of care has approximately an 85% probability of being cost effective using a willingness to pay threshold of £20,000 per QALY.

b. individual scenario analyses indicate that the cost effectiveness of bacteriophage therapy plus standard of care versus standard of care only is relatively stable, with the majority of scenarios estimating that bacteriophage therapy plus standard of care remains a dominant (that is, less costly and more effective) treatment strategy despite changes in key parameters.

c. the cost effectiveness of bacteriophage therapy plus standard of care was less clear when a number of conservative assumptions regarding its clinical effectiveness and cost were combined (for example, a 75% reduction in the probability of clinical resolution and a 200% increase in the cost of treatment).