Efficacy of a Travelers' Diarrhea Vaccine System in Travelers to India
A patch vaccine containing heat-labile toxin (LT) from enterotoxigenic Escherichia coli (ETEC) has demonstrated to be beneficial in reducing the rate and severity of travelers' diarrhea in Latin America. To evaluate the efficacy of this transdermal vaccine system in an area with a different diarrheal pathogen profile, an additional phase 2 study was conducted in European travelers to India.
For this multicenter, randomized, double-blinded, placebo-controlled field study 723 subjects were recruited; 603 (299 LT vaccine, 304 placebo) were included in the per-protocol-population (PPP).
Although the LT patch induced a measurable LT immune response in recipients, it failed to protect against LT ETEC or all-cause diarrhea. In the PPP the incidence rate of diarrhea as per primary endpoint was 6.0% (18 of 299) in the vaccine group and 5.9% (18 of 304) in the placebo group. Additionally, lower than expected rates of LT ETEC diarrheas were observed in India. The vaccine delivery system frequently produced rash and pruritus at the site of application, long term hyperpigmentation persisted in a minority of LT recipients, and also few site reactions were noted in the placebo group.
The evaluated patch vaccine failed to satisfy mainly with respect to protective efficacy. Noninvasive prophylactic agents against travelers' diarrhea, particularly vaccines against the most frequent pathogens, thus continue to be badly needed.
Traveler's diarrhea (TD), usually associated with additional symptoms, is still the most frequent illness among visitors from industrialized to low-income countries. Sequelae such as new onset postinfectious irritable bowel syndrome (IBS) are a recognized long-term complication. Enterotoxigenic Escherichia coli (ETEC) is the most frequent pathogen associated with TD. Heat-labile toxin (LT) and stable toxin (ST) genes are carried on plasmids in most ETEC strains and, when present, may be expressed either singularly (LT or ST) or in combination (LT and ST). The relative proportions of LT, ST, and LT/ST toxin-producing ETEC may vary across geographic areas. However, large surveys suggest that the overall toxin distribution among wild-type strains of ETEC is approximately 25% LT, 46% ST, and 29% LT/ST.
Antibodies to LT may block the secretory effects of LT and impair effective colonization by ETEC or other organisms, thus reducing the possibility of developing TD. A transcutaneous toxin-based vaccine system, which relies on LT antigen delivered via a patch, has been developed for the prevention of TD. This system consists of a self-adhesive patch containing LT manufactured by Intercell (Gaithersburg, MD, USA; Vienna, Austria), and a single-use device used to prepare the dermis prior to the patch administration. This route allows LT protein to exploit potent immunogenic Langerhans cells via the nearest draining lymph node,[6-9] while avoiding systemic toxicities associated with LT when administered via the oral, intranasal, or intrarectal routes. Encouraging results were seen in a phase 2 field trial—a 75% protective efficacy (PE) against moderate-to-severe diarrhea (defined as greater than four loose stools daily) and an 84% PE against severe diarrhea (six or more loose stools daily) were observed in travelers from the United States to Mexico and Guatemala, while the results in patients with ETEC diarrhea had been nonsignificant. Besides reducing the incidence, the LT patch vaccination also showed reduced severity of the diarrheal episode. Thus, it was decided to evaluate the efficacy of this TD vaccine system in travelers to Asia, an area with a different TD pathogen profile compared with Central America.
Study Design and Procedures
A multicenter, randomized, double-blinded, placebo-controlled field study was designed in accordance with the Code of Ethics of the World Medical Association (Declaration of Helsinki) and approved by the relevant Ethical Committees in the countries of origin (CO) and destination country, India. Eligible male and female subjects aged 18 to 64 years were recruited in the UK (five sites) and Germany (two sites). Preexisting plans for traveling to India were a prerequisite for participation at the sites in Germany, while in the UK a free holiday was offered. Between December 2009 and October 2010 (recruitment period), potential subjects were prescreened via telephone or Internet for eligibility criteria, medical history, and travel plans. The main exclusion criteria were current or significant chronic illness, including IBS, pregnancy or nursing, history of immunosuppression or deficiency, history of travelers' diarrhea or illicit drug use within the past year, known allergies to vaccine components, including adhesives, as well as drugs that may interfere with clinical diarrhea including enteric vaccines, antibiotics, antacids and antidiarrheals. Prior to performing study-related procedures at the first clinical visit, written informed consent was obtained both in the CO and later again upon arrival in India. After a brief clinical examination and a blood-draw for baseline LT-antibody levels, two transcutaneous immunizations (skin preparation with the Skin Preparation System buffer and LT or placebo single-use patch application) were administered 14 days apart on alternating deltoid regions by a clinician in the CO. Subjects received 37.5 µg LT or placebo patches as previously described according to randomized (1 : 1) group assignment. Patches were worn for 6 hours, prior to removal by the subject. Group assignments remained the same throughout the duration of the study. Assessment for adverse events (AEs) was performed 2 to 3 days after each vaccination by telephone interviews. In addition, subjects were instructed to maintain a daily record of local (patch site) and systemic AEs experienced during the vaccination and surveillance phase of the study using the sponsor-provided Vaccination Phase Diary. The clinician used the diary as a tool at clinical visits and discussed the recorded information with the subject in order to make accurate assessments. The Vaccination Phase Diary was collected at the “India Check-in” visit.
While the intent-to-treat (ITT) population included all subjects randomized, the modified ITT (mITT) describes all who traveled to India. Seven to 30 days after the second vaccination, subjects were expected to travel to any of the five sites there (North and South Goa, Delhi, Varanasi, Kolkata; option to visit several centers), but no further constraints had been made regarding travel routes, lodging and/or meals. Subjects were required to attend the “India Check-in” visit within 2 days of arrival in India and to remain there for a minimum of 7 days to complete the “Study Progress Visit”, but could return to their CO or stay in India for the “End of Surveillance Phase” visit. The subsequent surveillance phase lasted 17 days even if the subject returned to the CO during this interval. A blood sample was obtained for immunogenicity analyses and a Surveillance Phase Diary was provided at the India Check-in visit and subjects were instructed to capture daily stool incidence (formed and unformed). If the subject experienced a diarrheal episode, specific gastrointestinal (GI) symptoms (abdominal pain, fecal urgency, nausea, and vomiting) and interference with daily activities due to illness were recorded. Efficacy assessments were based on subject's diary-reported diarrheal events and laboratory analysis of stool samples submitted during the surveillance phase for detection of ETEC and other pathogens. Stool samples collected during diarrheal episodes were brought to the study clinic or study laboratory within 8 hours by the subject or via a stool courier system specifically established for the trial. The Surveillance Phase Diary was collected at the End of Surveillance Phase visit.
Similarly to the Vaccination Phase Diary, subject-assessed local and systemic AEs were also captured as well as any changes to the subject's concomitant medications during participation in the study including specific antibiotic use information.
In the follow-up phase, subjects were followed up for an additional 6 months and completed a final study visit where blood was drawn for evaluation of immunogenicity. Safety was assessed via AE review and concomitant medication review, and a patch site examination was also completed.
Definitions and Endpoints
A diarrhea episode was defined as three or more unformed stools in a 24-hour period, mild being three, moderate four to five and severe six or more bowel movements. The end of such an episode was the time of the last unformed stool that preceded a 48-hour period, during which no or only formed stools have been passed. Interference with daily activities meant that because of the illness the subject was confined to his/her room, changed plans, or had to seek medical advice and/or treatment. The primary endpoint of this study was defined as a case of moderate/severe diarrhea in which LT, LT and ST, or ST toxins (ETEC) were detected by either polymerase chain reaction (PCR) or DNA hybridization (and no copathogen was detected) from diarrheal stool samples that were collected during the first diarrheal episode. Secondary efficacy endpoints evaluated the following through the surveillance phase of the study: the incidence of moderate/severe all-cause (with or without ETEC diagnosis) diarrheal episodes, incidence of moderate/severe ETEC diarrheal episodes with or without co-pathogen, total unformed stool frequency from all (mild, moderate, or severe) all-cause diarrheal episodes, and total duration per subject of all (mild, moderate, or severe) all-cause diarrheal episodes. Tertiary efficacy endpoints based on characterization of all-cause diarrheal episodes through the surveillance phase of the study evaluated the following: interference with daily activities, loss of days, incidence of antibiotic treatment, and incidence of unscheduled clinical visits for medical treatment.
Immunogenicity and safety were evaluated as tertiary endpoints. Immunogenicity was based on serum anti-LT IgG, IgA, and toxin neutralizing antibody (TNA) titers from blood samples collected prior to first vaccination, upon entry in India, and at the final study visit. Geometric mean titers (GMTs), geometric mean fold ratios (GMFRs), and seroconversion rates for each serological parameter were calculated. A twofold rise or more of LT IgG or TNA titer relative to baseline was considered significant and defined seroconversion for LT IgG and TNA, respectively. A fourfold rise or more of LT IgA titer relative to baseline was considered significant and defined seroconversion for LT IgA. Safety evaluations were based on the subject Vaccination Phase Diary and Surveillance Phase Diary and on investigator-assessed AEs at all clinic visits.
For calculating a sample size, an attack rate of 10% ETEC moderate to severe diarrheal episode (as per primary endpoint definition) in the placebo group was estimated and a vaccine efficacy (VE) of 65% was derived from the previous field efficacy study. The sample size was calculated using 90% power and an alpha value of 0.05 for two-sided analyses. Using a 1 : 1 ratio of vaccine : placebo, a total per protocol study population of 622 subjects was computed; 716 participants were deemed necessary to account for an assumed dropout rate of 15%.
All primary, secondary, and tertiary efficacy-associated endpoints were analyzed using the per-protocol-population (PPP). VE was defined as [(attack rate in LT patch)/(attack rate in placebo)] × 100%. Chi-square or Fisher's exact tests were used to compare frequencies of binary variables such as occurrence of diarrhea illness, ETEC diarrhea, and diarrhea severity. After log-0 transformation of serological data, geometric mean titers and fold ratios were compared by t-tests. A p-value of 0.05 was accepted as statistically significant for two-sided comparisons.
The trial was registered with ClinicalTrials.gov, number NCT01040325.
The ITT population consisted of 723 subjects (363 LT vaccine group and 360 placebo), 722 received at least one vaccination, 691 both doses, and 603 (299 LT vaccine and 304 placebo) were included in the PPP. The mean age was 32.2 years (SD 10.4), the male : female ratio was 53 : 47; 87% were Caucasian, 10% Asian, the remaining 3% Black or others. No relevant differences between treatment groups were recorded in the demographic characteristics, concomitant medication used, other vaccines received, country of origin, or destination city in India.
In the PPP any diarrheal episode in the 17-day surveillance period (minimum stay 7 days) was experienced by 63 of 299 (21.1%) in the LT vaccine group and 61 of 304 (20.1%) in the placebo group (n.s.). The incidence rates of moderate to severe diarrhea and the proportion of ETEC is shown in Table 1. Co-pathogens detected with ETEC included Enteroaggregative E coli, Salmonella spp., Aeromonas spp., Entamoeba histolytica, Giardia lamblia, and Norovirus, each in less than 2% in both treatment groups. The same bacteria and viruses were detected in diarrheal episodes negative for ETEC, with enteroaggregative E coli (EAEC) being the most frequent (2.3%) in the LT vaccine group.
Table 1. Epidemiology, etiology, efficacy, and safety results
|Diarrheal incidence rates (% with 95% CI)|
|Moderate to severe, all causes (PPP)||12.7 (9.2, 17.0)||16.4 (12.5, 21.1)||n.s.|
|Moderate to severe, all causes (mITT)||11.9 (8.7, 15.9)||16.3 (12.5, 20.6)||n.s.|
|Moderate to severe, ETEC ± co-pathogen (PPP)||8.0 (5.0, 11.7)||8.2 (5.4, 11.9)||n.s.|
|ETEC diarrheal episodes with or without co-pathogen (PPP)||11.0||9.2||n.s.|
|Impact of travelers' diarrhea (%)|
|Mean total unformed stool frequency, all cause diarrhea (PPP)||10.7||13.0||n.s.|
|Loss of days due to diarrheal episode (PPP)||11.4||14.8||n.s.|
|Antibiotic treatment for all cause diarrhea (PPP)||3.7||7.9||p = 0.04|
|Any AE (%)||97.8||90.6||n.s.|
|Local AE incidence in number of subjects||4237 in 349||555 in 232||p < 0.0001|
|First vaccine application||89||48||p < 0.0001|
|Second vaccine application||93||40||p < 0.0001|
|First vaccine application||87||10||p < 0.0001|
|Second vaccine application||90||10||p < 0.0001|
|First vaccine application||80||8||p < 0.0001|
|Second vaccine application||79||3||p < 0.0001|
|First vaccine application||49||2||p < 0.0001|
|Second vaccine application||26||<1|| |
|First vaccine application||26||5||p < 0.0001|
|Second vaccine application||17||2||p < 0.0001|
|Local AE with grade ≥3 severity (n, %)||75, 21%||43, 12%||p = 0.002|
|Systemic AE incidence in number of subjects||1890 in 297||1831 in 231||n.s.|
|Systemic AE with grade ≥3 severity (n, %)||52, 14%||65, 18%||n.s.|
In the PPP the incidence rate of diarrhea as per primary endpoint was 6.0% (18 of 299) in the vaccine group and 5.9% (18 of 304) in the placebo group, while in the modified intent-to-treat (mITT) population these rates were 5.7 and 5.6%, respectively. Thus, the VE was near to zero (p = 1.0000). No difference was observed in the incidence of moderate or severe ETEC episodes with or without co-pathogen, and in the subgroup analysis of ETEC episodes with co-pathogen. No differences were observed between treatment groups in total duration of all all-cause diarrheal episodes, interference with daily activities, incidence of GI symptoms in subjects with all-cause diarrhea, or time-to-onset of first diarrheal episode. Additional efficacy results are summarized in Table 1. There was a statistically insignificant reduction of the total unformed stool frequency in the LT group compared with the placebo group (10.7 vs 13.0) with a difference of 2.4 stools [95% confidence interval (CI) −1.1, 5.8] per subject between treatment groups.
Immunogenicity assessment of the LT Vaccine System confirmed delivery of LT in the vaccinated group: While at baseline LT IgG GMT values were similar (LT: 812.7 and placebo: 840.2), it had increased in the LT group to 6408.7 as compared with 543.2 in the placebo group at the India Check-in Visit. The LT IgG GMFR was 7.89 (vs 0.65 in the placebo group) and 90.0% of subjects had seroconverted (vs 1.3% in the placebo group). The India Check-in Visit LT IgA GMT was also higher in the LT group (433.56) as compared with the placebo group (87.26) with a GMFR of 5.13 (vs 1.00 in the placebo group) and a seroconversion rate of 60.5% (compared with 0.0% in the placebo group). A similar response pattern was reported for TNAs with a GMT of 47.81, GMFR of 2.38, seroconversion rate of 58.5% in the LT group, and no response above baseline in the placebo group (GMT: 20.14, GMFR: 1.00, and 0.7% seroconversion). Anti-LT titers persisted in the LT group to the 6 months postvaccination Safety Follow-up Visit for LT IgG (GMT: 3457.54, GMFR: 4.22, and seroconversion: 80%) and LT IgA (GMT: 116.72, GMFR: 1.39, and seroconversion: 9.3%), and to a lesser extent for TNA (GMT: 29.83, GMFR: 1.47, and seroconversion: 32.6%).
In both trial groups more than 90% reported any AE (Table 1). Solicited local AEs of erythema, rash, pruritus, hyperpigmentation, and pain were reported at a significantly higher incidence in the LT group than in the placebo group. Almost 90% of LT recipients developed erythema and rash, while hyperpigmentation and pain were less frequent (Table 1). In the placebo group, erythema was commonly reported, but rash, pruritus, and hyperpigmentation occurred in 10% or less. Mild observed graded local reactions in the LT group were again common in around half the subjects (Table 1). Up to 14% were graded as severe local reaction, measuring >8 cm2. At the 6-month visit, 19% of LT group had persisting pigmentation at the site of the patch, and a single lesion in a placebo recipient was noted.
There were no apparent differences in incidence of the solicited systemic AEs following vaccination 1 in the two treatment groups, but there was a higher incidence of headache (p = 0.0275), malaise (p = 0.0014), and fever (p = 0.0231) following vaccination 2 in the LT group. The majority of these events were mild or moderate in severity, with slightly higher incidence of moderate to severe events in the LT group.
Essentially the transcutaneous patch vaccine failed in the primary and most secondary outcome targets. The immunogenicity to LT in subjects showed that the LT antigen could be delivered through the transdermal route. The reasons why the outcomes of this phase II and the phase III study, conducted in Latin America, are at odds with the original study by Frech need to be explored. The incidence of ETEC diarrheal episodes was approximately 5.9% (10% in the Frech study) overall, with only 1% positive for LT-secreting ETEC alone (5% in Frech Study). This was well below the expected 10%; the study was underpowered. To evaluate the primary endpoint on ETEC or specifically LT-ETEC episodes both studies showed no significant protection. While in this study there was not even a trend toward protection, the earlier study had shown a PE of 29% (95% CI −13 to 72; 15% vs 22%, p = 0.32). Similar findings are reported in the larger phase III study following the same design and delivery system in travelers to Latin America. The main difference in the two phase II studies is in rates of moderate to severe all-cause diarrhea. Frech reported a PE of 75% (95% CI 48–103; 5% vs 21%, p = 0.007) while we identified a PE of 21.4% (95% CI −17 to 47; n = 38 vs 50). The phase III Latin America study had similar numbers of cases in each arm with a PE of −1%. Neither of the new studies showed significant protection against all-cause, or moderate to severe ETEC diarrhea. Therefore the biological plausibility of the original protection reported by Frech is questionable, as protection was not achieved against the targeted pathogen but was noted against all diarrhea. The India and Latin America studies identified a small (2.2 stools per subject) reduction in frequency of stools.
Transdermally induced serum antibodies, either IgG or IgA, did not protect travelers to India, but in the Frech study, IgA conversion rates had been higher (78%) than the 60% who had seroconverted when measured on arrival in the CO. However, no correlation was observed in the primary endpoint incidence when LT recipients were stratified by high, medium, and low anti-LT responders.
It is unlikely that lack of sensitivity or inconsistency in test methodologies were major factors in the observed results for two reasons. First, results comparing LT and ST toxin detection in 20 versus 10 colony isolates by DNA hybridization produced identical results in 98% of samples analyzed and did not vary based on the type of toxin detected (tertiary endpoint). Second, DNA hybridization and PCR tests for concordance in toxin identification revealed concordance in 68% of all moderate/severe diarrheal episode cases evaluated, with no statistical differences noted.
The patch resulted in over 90% of subjects developing a local reaction, significantly more in the LT than in the placebo recipients. The local reactions were confirmed by the clinic staff and with rash graded as severe in 16% of subjects. The reactions included hyperpigmentation, which was still present in 19% at the 6-month review. The dermatitis resulted in a significantly increased use of topical anti-pruritic emolients and topical steroid agents during the vaccination phase by 28 LT recipients (7.7%), compared with no local treatment by the placebo group. This suggests that delivering protein molecules through the skin is associated with local acute inflammation and long lasting hyperpigmentation, which may well limit the acceptability by subjects.
Although this vaccine trial failed, there is a continuing need to protect travelers against diarrhea at high-risk destinations. The focus on preventing a single pathogen in what is otherwise a syndrome of TD may ultimately not be an ideal approach in future vaccine studies, where combination antigen vaccines might produce greater success in reducing morbidity. Even a vaccine preventing just a third of the cases would be welcome. Not only would it reduce incapacitation and frustration at times of high expectations—both in leisure and professional travel—but hopefully it might also diminish the small risk of subsequently experiencing IBS. A future vaccine would need to be safe, as tolerance even for moderate AEs is very limited among healthy travelers. Noninvasive routes have the potential for self-administration, which for travelers would be very attractive and lead to much wider usage. New delivery routes need to have minimal local reactivity and appropriate antigens would need to be selected with this in mind. These novel routes should be encouraged to improve uptake and reduce costs.
The authors thank all colleagues who contributed to the conduct of this study, in particular former employees from Intercell: Sarah Frech and Gregory Glenn for contribution to study design, Astrid Kaltenboeck, Amanda Moore, and Rosalind Hollingsworth for operational contributions to study, Bo Anderson for serological testing and Christoph Klade for review of the study results.
R. S., R. B., and J. P. C. drafted and finalized the manuscript. All authors were invited to comment on the initial draft and they approved with the final text.
Declaration of Interests
All authors were investigators in this study funded by Intercell AG and received research sponsorship for this project, except for S. D. and K. W., who were employees of this vaccine producer (now merged into Valneva SE). R. S. currently is Principal Investigator in a traveler's diarrhea therapy study sponsored by Dr Falk Pharma, and R. B. served on the Advisory Board of Norgine Pharmaceutical. None of the other authors have additional conflicts of interest relevant to this study to declare.