NSAID use and unnatural deaths after cancer diagnosis: a nationwide cohort study in Sweden

BMC Cancer

Table 3 Association of NSAID use with risk of death due to suicide or accident after cancer diagnosis, separate analysis by status of previous medication use, a cohort study of 388,443 cancer patients diagnosed between Oct 2005 and Dec 2014 in Sweden

Characteristics	Completed suicide			Death due to accident
	N	Crude IR (per 1000 person-years)	HR (95% CI)	N	Crude IR (per 1000 person-years)	HR (95% CI)
Low-dose aspirin
Without previous use
Off medication (Ref) ^a	187	0.22	1.0	1060	1.22	1.0
On medication (first-time use) ^b	10	0.53	1.83 (0.84–3.96)	60	3.20	1.36 (1.03–1.80)
With previous use
Off medication (Ref) ^a	41	0.30	1.0	567	4.20	1.0
On medication (repeated use) ^b	49	0.23	0.69 (0.42–1.13)	591	2.77	0.58 (0.51–0.66)
	P for interaction < 0.001			P for interaction < 0.001
Non-aspirin NSAIDs
Without previous use
Off medication (Ref) ^a	128	0.25	1.0	1233	2.38	1.0
On medication (first-time use) ^b	3	0.46	0.91 (0.13–6.55)	19	2.94	1.81 (1.10–2.96)
With previous use
Off medication (Ref) ^a	146	0.21	1.0	986	1.45	1.0
On medication (repeated use) ^b	10	0.33	0.92 (0.37–2.29)	40	1.30	0.77 (0.52–1.13)
	P for interaction = 0.99			P for interaction = 0.008

Abbreviations: NSAID, non-steroidal anti-inflammatory drugs; IR, incidence rate; HR, hazard ratio; CI, confidence interval
^a Off medication time included follow-up time accumulated among patients without any dispensed NSAIDs during follow-up, as well as the non-medicated periods from patients that had any dispensed NSAIDs
^b Analyses were adjusted for sex, age at diagnosis, calendar year of cancer diagnosis, highest educational level, occupation, cohabitation status, cancer stage, cancer type, Chronic Disease Score and history of psychiatric disorders; time since cancer diagnosis was used as the underlying timescale

ISSN: 1471-2407