Table 2.

Evidence

Author, year	Study design	Location	Extended method continuation	Loss to follow‐up	Implant type	Duration	Pregnancy outcomes and assessment	Strengths	Weaknesses	Quality
McNicholas et al. 201715	Noncomparative cohort	USA	n=223 completing 3rd year	5% at 24 mo	1 rod ENG	Up to 2 y extended use	Year 4: Zero pregnancies per 100 woman years	Assessment of ancillary methods of contraception (barrier methods etc.) and comorbidities that may affect fertility (BMI, age)	Small sample size completing each additional year	II‐3, Fair
			n=102 completing 4th year				Year 5: Zero pregnancies per 100 woman years	Detailed description of study population provided	Self‐selected cohort extending duration
							Assessment: Self‐report (with validation through medical record review where possible)	Study results generalizable with diverse demographics of study population	Pregnancy assessment – validation only where possible
Ali et al. 201620	Noncomparative cohort analysis from RCT	Brazil, Chile, Dominican Republic, Hungary, Thailand, Turkey, Zimbabwe	n=311 completing 4th year	2%	1 rod ENG	Up to 2 y extended use	Year 4: Zero pregnancies per 100 woman years	Detailed description of study population provided	Small sample size completing each additional year	II‐3, Fair
			n=204 completing 5th year				Year 5: Zero pregnancies per 100 woman years	Study results generalizable with multiple study sites	High implant discontinuation rate over extended duration (48%)
							Assessment: Urine pregnancy testing if suspicion of pregnancy	Pregnancy assessment well defined
								Low loss to follow‐up
Yao and Du, 200321	Nonrandomized, comparative, prospective observational	Not discussed	n=51 completing the 4th year	0%	1 rod ENG	Up to 1 y extended use	Zero pregnancies	Unable to assess generalizability	Small sample size completing each additional year	II‐3, Poor
							Assessment: No information		No assessment for potential confounders
									Self‐selected cohort extending duration
									No comparison group
Sivin et al. 200116	Multicenter, pharmacokinetic trial	USA	n=43 completing 6th year	No information	2 rod LNG	Up to 2 y extended use	2 pregnancies in the 5th year and 3 in the 7th year	Multiple study sites	Small sample size completing each additional year	II‐3, Poor
		non‐USA	n=39 completing 7th year				Assessment: No information		Self‐selected cohort extending duration
									No comparison group
Zheng et al. 199922	Noncomparative, multicenter prospective observational	China	n=151 completing 4th year	1%	1 rod ENG	Up to 1 y extended use	Zero pregnancies	Assessment of appropriate confounders including: age (age at enrollment was <35 y old)	Small sample size completing each additional year	II‐3, Poor
							Assessment: Urine pregnancy testing if suspicion of pregnancy	Pregnancy assessment well defined	Self‐selected cohort extending duration
									No comparison group
									Lack of generalizability secondary to limited demographic data and all study sites within one country
Kiriwat et al. 199823	Noncomparative cohort	Thailand	n=47 completing 4th year	8%	1 rod ENG	Up to 1 y extended use	Zero pregnancies	Pregnancy assessment well defined	Small sample size completing each additional year	II‐3, Poor
							Assessment: Urine pregnancy testing if suspicion of pregnancy		No assessment for potential confounders
									Self‐selected cohort extending duration
									No comparison group
									Lack of generalizability secondary to limited demographic data

Abbreviations: ENG, Etonogestrel; LNG, Levonorgestrel.