Table 2 Strengths and limitations of epidemiologic study designs for examining the associations between the gastrointestinal microbiome and prostate cancer risk

Case–control:

● Quick

● Cannot truly establish temporality or differentiate between cancer-induced and pre-cancerous changes in the GI microbiome

Diagnostically-confirmed latent and invasive prostate cancer cases compared to each other and to matched controls

● Relatively inexpensive

● No follow-up required

● Difficult to obtain appropriate control group

Prospective Cohort:

● Ability to assess changes from multiple samples as individuals transition from healthy to cancerous state

● Need very large sample size to be able to obtain enough incident prostate cancer cases

Large cohort of older men followed over time with regular assessments of GI microbiome and prostate cancer status

● Need long follow up time

● Can obtain data on incident cases

● Extremely expensive

● Potential biases due to loss to follow

● Can continue to obtain data throughout course of treatment and progression to assess post-diagnostic longitudinal and treatment-related changes

● Can evaluate mortality as an end point

Retrospective Cohort:

● Likely to have clearer temporality between assessment of microbial/metabolic profile and prostate cancer development than in a case–control study

● Requires availability of previously collected and appropriately preserved samples (or data)

Previously collected samples on a large cohort of men, who were healthy at baseline, assessed for current prostate cancer status

● Participants must have been cancer-free at time of sample collection

● Less expensive than prospective cohort study

● Multiple longitudinal samples are unlikely to be available