We conducted an open-label 5-arm randomized controlled trial from 27 November 2012 to 30 November 2013 in Mirpur, an urban neighborhood in Dhaka, Bangladesh. The trial enrolled participants from 5 different sections of Mirpur. During the duration of the trial no polio vaccination campaigns were conducted in or around the study site. Infants were assigned randomly to one of five arms using a block randomization scheme of 65 blocks with a block size of 18 and an allocation ratio of 4:4:3:3:4 (Fig. 1). The tOPV arm received tOPV at ages 6, 10 and 14 weeks; the bOPV arm received bOPV at ages 6, 10 and 14 weeks; the IPV arm received IPV at ages 6 and 14 weeks; the f-IPV arm received f-IPV at ages 6 and 14 weeks; and the f-IPV/bOPV arm received f-IPV at ages 6 and 14 weeks with bOPV at age 10 weeks. All participants received tOPV at age 18 weeks (Table 1 in Supplementary Appendix).

The study’s three primary objectives were to compare immunogenicity of (1) f-IPV and bOPV with bOPV alone; (2) 3 doses of tOPV with 3 doses of bOPV; and (3) 2 doses of intramuscular IPV with 2 doses of f-IPV.

Infants were recruited at age 6–7 weeks (42–51 days), if the parents were willing to participate, comply with study procedures, and provide written informed consent. Exclusion criteria included (1) receipt of any polio vaccine before enrollment; (2) diagnosis or suspicion of immunodeficiency or a bleeding disorder; (3) known allergy to polio vaccines or constituents; (4) any acute illness such as vomiting, diarrhea or infection immediately before enrollment; and (5) an infant who was part of a multiple birth. Enrolled participants were withdrawn from the study if requested by their parents or if they received polio vaccine outside of the study.

Study physicians administered all study vaccines and routine non-polio vaccines for infants as recommended by the Bangladesh Ministry of Health and Family Welfare. Intramuscular IPV (0.5 ml) was administered using a standard needle and syringe. Intradermal f-IPV (0.1 ml) was administered using NanoPass MicronJet 600 (MJ600), a microneedle device with three microneedles (0.6 mm in length) that attaches to an intradermal syringe. Multiple clinical trials have been conducted using MJ600 [9–12]. IPV and f-IPV were administered in the anterolateral thigh, opposite the side used for routine immunization of injectable vaccines.

Blood samples (1 ml) were obtained by venipuncture at ages 6, 14, and 18 weeks from all participants and at age 15 weeks from participants assigned to IPV or f-IPV arms before administering any scheduled study vaccine. Sera were stored at −20 °C and tested for antibodies to poliovirus types 1, 2 and 3 at the Centers for Disease Control and Prevention (CDC), Atlanta, USA using microneutralization assay. Titers below a dilution of 1:8 were considered negative for presence of poliovirus antibodies and the highest measurable titer was 1:1448. Parents were asked to collect a stool specimen (8 g) from participants 1 week after tOPV administration at age 18 weeks. Stool specimens were stored at −20 °C and tested at CDC for presence of poliovirus by type [13].

No published studies were found to have administered f-IPV with bOPV, or 3 doses of bOPV. Therefore, for sample size calculations based on limited evidence, we assumed seroconversions of 85% for types 1 and 3 with f-IPV and bOPV and 95% with 3 doses of bOPV [14,15]. For tOPV, we assumed sero-conversions of 75% for type 1 and 65% for type 3 [16]. Therefore, a sample size of 207 per arm would be sufficient to obtain a power of 90% with two-sided α of 0.05 to detect a difference in seroconversion of at least 10% when comparing 3 doses of bOPV with 3 doses of tOPV, and 2 doses of f-IPV and 1 dose of bOPV with 3 doses of bOPV. No published studies were found to have reported immunogenicity of IPV or f-IPV with two doses 8 weeks apart at ages 6 and 14 weeks. For a non-inferiority comparison, we assumed a sero-conversion of 90% with both IPV and f-IPV with a non-inferiority margin of 10% [14,17]. For this comparison a sample size of 155 per arm is required for a power of 90% with one-sided ˛ of 0.05. Hence, the effective sample size for the trial was 931, with an enrollment target of 1170 assuming 20% attrition (Table 1 in Supplementary Appendix).

Seroconversion was defined as either conversion from seronegative to seropositive or a four-fold increase in antibody titers between two specimens after adjusting for decay of maternal antibodies. The half-life of maternal antibodies was assumed to be 28 days [14,18]. The primary analytical approach was intent-to-treat for participants with serological results. The primary end-point was seroconversion at age 18 weeks. To compare immunogenicity across study arms, the proportion of participants who sero-converted were compared using Fisher’s exact test (two-tailed). Priming was defined as a seroconversion response at age 15 weeks after receipt of the second IPV/f-IPV dose among those who did not seroconvert by age 14 weeks after one IPV/f-IPV dose at age 6 weeks. Reverse cumulative distribution curves, which are constructed by representing on the vertical axis the percent of subjects with antibody titers equal to or greater than that marked in x-axis, were used to compare distribution of antibody titers by study arms [19].

The study protocol was reviewed by icddr,b’s Institutional Review Board (IRB). The study was conducted in compliance with good clinical practice guidelines. UNICEF assisted in the procurement of vaccines used in this study. OPV was manufactured by Sanofi Pasteur and IPV was manufactured by the Netherlands Vaccine Institute (NVI). NanoPass Technologies Ltd. donated the supplies of MJ600. UNICEF, Sanofi Pasteur, NanoPass, and NVI had no role in the study design, implementation, data analysis, or interpretation of study results. The study was registered with Clinicaltrials.gov (NCT01813604). Adverse events data were reviewed by the Data Safety Monitoring Board (DSMB) of icddr,b.

The study was funded by the Global Immunization Division of the Centers for Disease Control and Prevention. CDC staff participated in the study design, sample testing, data analysis and decision to submit for publication.

The study enrolled and randomized 975 participants and of these, 922 (95%) with blood specimens available at ages 6 and 18 weeks were included in the primary end-point analysis (Fig. 1). Enrollment was stopped after enrolling 975 participants as the study had achieved its effective sample size due to lower than anticipated study attrition. No statistically significant differences were observed at baseline among participants who completed the study compared with those who did not (data not shown) except that median type 2 antibody titers at baseline were lower for those who completed the study (1:28 vs 1:41, Kruskal–Wallis = 0.036). No other significant differences in baseline characteristics, including seroprevalence to polioviruses, were observed among study arms (Table 1).

The median bleb diameter after intradermal injection with MJ600 was 10 mm and 99% of the participants had no residual liquid present on the skin following the injection.

Seroconversion to poliovirus type 1 (PV1) after 2 and 3 doses was higher in the bOPV arm compared with the tOPV arm (2 doses: 93% vs 87%, p = 0.047; 3 doses: 99% vs 94%, p = 0.019; Table 2). PV1 sero-conversion with 2 doses of IPV (95%) was statistically no different from that observed with 3 doses of tOPV or bOPV. PV1 seroconversion with 2 doses of f-IPV and 1 dose of bOPV was higher than that observed with 2 doses of f-IPV alone (p = 0.005) and no different from that with 3 doses of tOPV or bOPV.

Seroconversion at 18 weeks to PV2 was higher with 3 doses of tOPV compared with 2 doses of IPV p = 0.002) or f-IPV in either f-IPV arms (p < 0.001). Seroconversion to PV3 was statistically no different with 3 doses of tOPV (95%) compared with 3 doses of bOPV (94%), 2 doses of IPV (97%) or f-IPV (89%), or 2 doses of f-IPV with 1 dose of bOPV (94%).

Compared with IPV, f-IPV failed the non-inferiority test for all serotypes for seroconversion observed with 1 or 2 doses (Fig. 2). Additionally, compared with IPV, f-IPV failed the non-inferiority test for all serotypes for priming response observed at 15 weeks.

Reverse cumulative distribution curves for antibody titers by study arm at age 18 weeks show that the highest titers were reached for PV1 in the bOPV arm, PV2 in the tOPV arm and PV3 in the IPV arm (Fig. 3). f-IPV was associated with the lowest titers for all three poliovirus types among those receiving type specific vaccines. One dose of IPV or f-IPV was not associated with a substantial change in distribution of antibody titers, despite the high degree of priming with 1 dose; however, within a week of the second dose of IPV or f-IPV, a rapid rise in antibody titers was observed (Fig. 1 in Supplementary Appendix).

One week after receiving tOPV at age 18 weeks, 15%, 6%, and 8% of participants in the tOPV arm were excreting PV 1, 2, and 3, respectively (Table 2). Among participants in the bOPV arm, 61% were excreting PV2 1 week after receiving tOPV. The percent of participants excreting type 1 poliovirus was statistically lower in the bOPV arm compared with the f-IPV/bOPV arm (4% vs 13%, Fischer’s exact = 0.001). The percent excreting PV3 was statistically lower in the bOPV arm compared with f-IPV/bOPV arm (6% vs 14%, Fischer’s exact = 0.013). No statistically significant differences in percent excreting polioviruses by type were observed between IPV and f-IPV arms.

No adverse events (AE) were reported among participants 30 min after receiving the study vaccine. During follow-up (age 6–19 weeks), 68 AE were reported among participants; 11 were considered serious AE (SAE), including hospitalization or death (Table 2 in Supplementary Appendix). Three infants died during follow-up: two in the sequential f-IPV/bOPV arm and one in the f-IPV arm. No AE/SAE were attributed to trial vaccines or MJ600 by the DSMB.