Asturias 2007.

Methods	Study design: open‐label, randomised trial Setting: 3 well‐child public clinics, Guatemala city, Guatemala Study dates: April to November 2004
Participants	Sample size: 500 healthy, full‐term infants naive to polio vaccination were recruited from April to November 2004 Age: mean = 59.5 days Sex: male = 253, female = 247 Inclusion criteria: "Healthy, full‐term, 6–11‐week old infants attending 3 well‐child public clinics in Guatemala City were eligible" (quote) Exclusion criteria: "(1) received polio, hepatitis B (HB), Haemophilus influenzae type b (Hib), or diphtheria‐tetanus toxoids–pertussis (DTP) vaccines; (2) a history of any disease preventable by these vaccines; (3) a confirmed immunosuppressive condition; (4) received immunosuppressive drugs or blood‐derived products; (5) major congenital defects or serious chronic illness; (6) a history of any neurological disorders or seizures; or (7) allergies to any component of the vaccines." (quote)
Interventions	Group A (n = 166): IIII at 2, 4, 6 and 12 months Group B (n = 168): IIOO at 2, 4,6 and 12 months Group C (n = 166): OOOO at 2, 4, 6 and 12 months
Outcomes	Antibody titres to poliovirus type 1, 2 and 3 (geometric mean antibody titres) Poliovirus detection in stool Local reactions (tenderness, swelling, and redness). Systemic adverse events (fever, irritability, lethargy, and vomiting) Timing of outcome assessment: 2, 6, 7, 12, and 13 months
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Infants were randomized using a permuted block design of 6–12 at a 1:1:1 ratio within each study" Comment: probably done properly, since the table of baseline characteristics of included participants is balanced
Allocation concealment (selection bias)	Low risk	Quote: "Allocation to study group was done by opening sealed, sequentially numbered envelopes"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Comment: no blinding but performance is not likely to be influenced by lack of blinding in terms of intervention or co‐interventions
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Comment: no blinding. Objective outcome (antibody titres) is not likely to be influenced by lack of blinding. Diary cards and digital thermometers were used by parents to record adverse events. Diary cards were collected at the next scheduled visit, and the parent was interviewed to ensure completeness of the information. These outcomes could have an unclear risk of bias but they were not meta‐analysed.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Comment: Out of 500 infants enrolled, 444 (88.8%) were available for primary end‐point analysis, and 439 (87.8%) completed the last study visit.
Selective reporting (reporting bias)	Low risk	Comment: no evidence of selecting reporting
Other bias	Low risk	Comment: no evidence of other bias
Conflict of interest	Unclear risk	Financial support: financially supported by Sanofi Pasteur. Independent funds were used to support statistical analyses, and support for one of the authors was provided by a Fogarty International Research Scientist Development Award (grant KO1 TW006659).