Summary of findings 2. Hydromorphone compared with morphine for people with moderate to severe cancer pain.

Hydromorphone compared with morphine for people with moderate to severe cancer pain
Patient or population: people with moderate to severe cancer pain Setting: inpatients, outpatients and day patients Intervention: hydromorphone Comparison: morphine
Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with morphine	Risk with hydromorphone
Participant‐reported pain intensity (as measured by BPI and VAS) Follow‐up: 12 weeks	For pain intensity measured by BPI at 24 days, the results showed slightly higher mean endpoint scores for 'worst pain' in morphine group and similar mean scores for 'average' and 'least' pain in hydromorphone and morphine groups. For pain intensity measured by VAS from weeks 1–12, both morphine and hydromorphone groups had mean pain levels of 'no worse than mild pain.' Evidence of hydromorphone vs morphine on pain intensity was very uncertain.		—	433 (2 RCTs)	⊕⊝⊝⊝ Very lowa,b,c	1 study (n = 200) using subscale data derived from BPI scale (0–10; higher = worse outcome): mean endpoint score for 'worst pain:' hydromorphone 3.5 (SD 2.9, n = 99); morphine 4.3 (SD 3.0, n = 101). Mean scores on 'least pain' and 'average pain' were almost identical. 1 study (n = 233) using VAS scale (0–10; higher = worse outcome) measured pain intensity from week 1 to week 12 of treatment. Both groups had identical scores at all the measured timepoints (P > 0.05).
Participant‐reported pain relief Follow‐up: mean 12 weeks	Study population		RR 0.99 (0.84 to 1.18)	233 (1 RCT)	⊕⊝⊝⊝ Very lowa,b,c	—
	705 per 1000	698 per 1000 (593 to 832)
Specific adverse events – nausea Follow‐up: 24 days	Study population		RR 0.94 (0.66 to 1.30)	200 (1 RCT)	⊕⊝⊝⊝ Very lowa,b,c	—
	396 per 1000	372 per 1000 (261 to 515) fewer
Specific adverse events – vomiting Follow‐up: 24 days	Study population		RR 0.87 (0.58 to 1.31)	200 (1 RCT)	⊕⊝⊝⊝ Very lowa,b,c	—
	337 per 1000	293 per 1000 (195 to 441) fewer
Specific adverse events – dizziness Follow‐up: 24 days	Study population		RR 1.15 (0.71 to 1.88)	200 (1 RCT)	⊕⊝⊝⊝ Very lowa,b,c	—
	228 per 1000	262 per 1000 (162 to 428) more
Specific adverse events – constipation Follow‐up: 24 days to 12 weeks	The higher incidence of constipation of hydromorphone occurred at a shorter treatment point (at 24 days of treatment), but not a longer treatment point (12 weeks).		—	433 (2 RCTs)	⊕⊝⊝⊝ Very lowa,b,c	2 studies reported the incidence of constipation at different timepoints. 1 study (n = 200) measured at 24 days of treatment found a significantly higher incidence of constipation with hydromorphone than with morphine (RR 1.56, 95% CI 1.12 to 2.17; P = 0.009). 1 study (n = 233) measured at 12 weeks' follow‐up found no clear difference between 2 groups (RR 0.65, 95% CI 0.42 to 1.00; P = 0.055).
Quality of life	Not reported.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). BPI: Brief Pain Inventory; CI: confidence interval; n: number of participants; RCT: randomised controlled trial; RR: risk ratio; SD: standard deviation; VAS: visual analogue scale.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded twice due to very serious study limitations: all studies were rated at high risk of bias for at least two domains.
^bDowngraded once due to serious imprecision: studies contained fewer than 200 participants in each treatment arm, and sample size was smaller than optimal information size (Guyatt 2011); CI around estimate of effect was wide and included no effect and appreciable benefit/harm.
^cDecision made not to downgrade due to publication bias: although publication bias was highly suspected due to the small number of trials identified, this outcome had already been downgraded three times for other factors, therefore, further downgrading would be inappropriate.