Introduction
Human leptospirosis is a bacterial zoonotic infection that accounts for approximately one million cases and 60,000 deaths globally each year [Reference Costa, Hagan, Calcagno, Kane, Torgerson and Martinez-Silveira1]. Leptospirosis is caused by spirochete bacteria, Leptospira, which are most often carried by rodents, livestock, and dogs [Reference Adler and de la Pena Moctezuma2]. Leptospires are excreted via animal urine into the environment, where humans may become infected through contact with infected animals or contaminated soil and water [Reference Adler and de la Pena Moctezuma2]. An increasing number of leptospirosis outbreaks have been reported around the world, particularly following heavy rainfall and flooding [Reference Lau, Smythe, Craig and Weinstein3].
Human leptospirosis poses a significant public health risk in the Pacific Islands, with Oceania representing the highest burden of leptospirosis worldwide, as measured by morbidity (150.68 cases per 100,000 per year) and mortality (9.61 deaths per 100,000 per year) [Reference Costa, Hagan, Calcagno, Kane, Torgerson and Martinez-Silveira1]. Climate change presents an additional threat to leptospirosis burden, with flooding predicted to become more intense and frequent in the region [Reference Masson-Delmotte, Zhai, Pirani, Connors, Péan and Berger4]. However, there is currently limited knowledge about the primary drivers of human Leptospira infection in the Pacific Islands. The local incidence of leptospirosis remains poorly documented due to the unavailability of laboratory diagnosis, limited medical awareness, and non-specific symptoms that overlap with other tropical diseases [Reference Guernier, Goarant, Benschop and Lau5]. Consequently, it has been difficult for local authorities to implement mitigation strategies [Reference Guernier, Goarant, Benschop and Lau5].
Whilst an existing review examined the drivers of leptospirosis at a broader scale [Reference Mwachui, Crump, Hartskeerl, Zinsstag and Hattendorf6], many Pacific Island studies were excluded from their analysis. As a result, their synthesized findings are not necessarily applicable to the context of these vulnerable islands. By better understanding the local drivers of infection within the region, appropriate mitigation measures can be implemented to combat leptospirosis outbreaks.
Hence, the aim of this systematic review was to identify the drivers of human Leptospira infection in the Pacific Islands, which will lead to recommendations for public health interventions to reduce future disease burden. We have considered the Pacific Islands to refer to the three ethnogeographic groups of Melanesia, Micronesia, and Polynesia, which includes New Zealand but excludes neighbouring islands of Australia, Indonesia, and the Philippines [Reference Lal and Fortune7].
Materials and methods
This systematic review has been reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [Reference Page, McKenzie, Bossuyt, Boutron, Hoffmann and Mulrow8].
Search strategy
Ovid Medline (1946–present), Scopus (1960–present), Web of Science Core Collection (1900–present), and Ovid Embase (1974–present) electronic databases were searched in December 2022 using search terms in the title, author keyword, and abstract fields. The search also included subject headings in Embase and Medline (i.e. Emtree and Medical Subject Headings (MeSH), respectively), as well as indexed terms for Scopus (i.e. indexed keywords) and Web of Science Core Collection (i.e. Keywords Plus) (see Supplementary Material S1 for details). There were no language or date restrictions. The reference and citation lists (Web of Science Core Collection, Scopus, and Google Scholar) of included studies, as well as relevant excluded reviews, were also screened for additional studies. Reference and citation screening were performed independently by two reviewers (S.K. and J.S.) in June 2023.
Study selection
All identified articles were exported into Covidence, where duplicates were manually and independently removed by two reviewers (S.K. and J.S. for English language studies and S.K. and P.W. for French language studies). The reviewers then independently screened the articles by title and abstract, and full text, according to the eligibility criteria (see Table 1).
Eligibility criteria for study inclusion using the population, exposure, comparator, and outcomes (PECO) framework [Reference Morgan, Whaley, Thayer and Schunemann9]
a The following Pacific Islands or regions were included, representing Melanesia, Micronesia, and Polynesia: Fiji, New Caledonia, Papua New Guinea, Solomon Islands, and Vanuatu (Melanesia); Guam, Kiribati, Marshall Islands, Federated States of Micronesia, Nauru, Northern Mariana Islands, Palau, and the US minor outlying islands, which include Baker Island, Howland Island, Jarvis Island, Johnston Atoll, Kingman Reef, Midway Atoll, Palmyra Atoll, and Wake Island (Micronesia); and American Samoa, New Zealand, Cook Islands, Easter Island, Hawaii, Niue, Pitcairn, Samoa, Tokelau, Tonga, Tuvalu, Wallis, Futuna, and French Polynesia, which include the Society Islands, Tuamotu Archipelago, Gambier Islands, Marquesas Islands, and Austral Islands (Polynesia) [Reference Foster and West10].
b Hospitalization, mortality, and other possible outcomes or complications of Leptospira infection were not examined in this review.
c Grey literature, predictive models, reviews, conference abstracts, editorials, and book chapters were excluded.
Any discrepancies that arose were resolved through discussion and consensus decision between the two reviewers, with a third reviewer consulted when consensus was not reached.
Data extraction
For each included article, two reviewers (S.K. and J.S. for English language studies and S.K. and P.W. for French language studies) manually and independently extracted data into an Excel spreadsheet with the following headings: (a) citation details; (b) study location; (c) year(s) and month(s) of study data collection; (d) study design; (e) population and sample characteristics; (f) driver(s) assessed, how they were measured, and confounders considered; (g) leptospirosis outcome(s) and how they were measured; (h) serogroup(s) and method of identification; (i) statistical approaches; and (j) findings related to the association between the driver(s) and leptospirosis outcome(s).
Risk-of-bias assessment
Each study was allocated to a level of evidence according to the Australian National Health and Medical Research Council (NHMRC) evidence hierarchy for aetiological studies [11]. The NHMRC evidence hierarchy consists of levels of evidence from I to IV, where level IV represents the study design least robust at answering a research question. As ecological and test-negative case–control study designs are not included in the existing evidence hierarchy, we classified these studies as level IV.
The risk of bias for each study was further evaluated independently by two reviewers (S.K. and J.S. for English language studies and S.K. and P.W. for French language studies) using the validated ‘Methodological Evaluation of Observational Research Checklist’ (MEVORECH) tool for risk factor studies [Reference Shamliyan, Kane, Ansari, Raman, Berkman and Grant12] (see Supplementary Table S1 for details). Discrepancies were resolved through discussion and consensus decision between the two reviewers (S.K. and J.S.), with a third reviewer (P.W.) consulted when consensus was not reached.
Synthesis of results
After data extraction, drivers were classified into the following categories: sociodemographic, occupational, lifestyle, and environmental. Activities performed at a recognized occupational setting, such as an abattoir or farm, were classified as occupational, whereas activities conducted recreationally or at home, including backyard gardening/farming, were considered as lifestyle drivers of infection. Study characteristics and findings were synthesized descriptively, using tables and narrative descriptions. Findings were considered statistically significant if P < 0.05 or if the authors reported the finding as significant in studies where no P-value was reported. Both unadjusted and adjusted findings are reported for included studies; however, when reporting the number of studies with significant findings, the adjusted estimates have been used. Quantitative synthesis was not performed due to heterogeneity across studies with respect to populations, exposures, comparators, and outcomes.
Protocol registration
The review protocol was registered prospectively with International Prospective Register of Systematic Reviews (PROSPERO) (CRD42022360109) [Reference Kharwadkar, Stanhope and Weinstein13]. Modifications were made to the initial protocol prior to database searches to include the Pacific Islands described in Table 1.
Results
Study selection
There were 39 articles included in this review (see Figure 1). One of these articles reported four distinct studies on leptospirosis outbreaks between 1985 and 1986 in New Caledonia [Reference Brethes, Puech, Fraisse, Dubois, Domenech and Bourdin14], and hence, a total of 42 studies were included. Reference and citation screening of included articles and relevant excluded reviews yielded no further eligible articles.
PRISMA flow diagram [Reference Page, McKenzie, Bossuyt, Boutron, Hoffmann and Mulrow8] for the article selection process.
Study characteristics
Of the 42 included studies, the most common study design was cross-sectional (n = 17, 40%) followed by case series (n = 14, 33%) and test-negative case–control designs (n = 7, 17%), all of which were classified as level IV on the NHMRC evidence hierarchy. Three articles (7%) reported prospective cohort studies (level II) [15, Reference Pittavino, Dreyfus, Heuer, Benschop, Wilson and Collins-Emerson16, 30], and the remaining study was an ecological study (level IV) [Reference Mackintosh, Schollum, Blackmore and Marshall17] (see Table 2).
Study location, period, design, and population and sample characteristics of included studies
Abbreviations: IQR, interquartile range; NHMRC, National Health and Medical Research Council; NR, not reported.
a Reported the same study period from 1985 to 1986 in New Caledonia.
b Reported the same prospective cohort study conducted in New Zealand from February 2008 to May 2011. Reference [Reference Dreyfus, Benschop, Collins-Emerson, Wilson, Baker and Heuer29] reported one sample undertaken between November 2009 and March 2010.
c Study design was reported by the authors as case–control; however, this referred to the farms and not the individuals measured. For this review, the study design was classified as ecological as the exposures were measured after the outcome.
d Reported on the same seroprevalence study conducted in American Samoa in 2010.
e Reported on the same seroprevalence study conducted in Fiji in 2013.
In terms of geographic distribution, 16 studies (38%) were conducted in New Zealand and 24 studies (57%) represented other Pacific Islands, including New Caledonia (n = 8, 19%), Hawaii (n = 4, 10%), American Samoa (n = 3, 7%), Fiji (n = 3, 7%), French Polynesia (n = 2, 5%), Wallis/Futuna (n = 2, 5%), Palau (n = 1, 2%), Vanuatu (n = 1, 2%), and Federated States of Micronesia (n = 1, 2%), and the final study examined several Pacific Islands [Reference Berlioz-Arthaud, Kiedrzynski, Singh, Yvon, Roualen and Coudert20].
With respect to study outcomes, there were 23 studies (55%) reporting on human leptospirosis incidence, 18 seroprevalence surveys (43%), and two studies (5%) investigating both incidence and seroprevalence. Approximately half of the studies (48%) used a combination of direct diagnostic methods, such as polymerase chain reaction (PCR), culture and dark field microscopy, and indirect methods, such as microscopic agglutination test (MAT) and enzyme-linked immunosorbent assay (ELISA), to investigate Leptospira infection. There were 21 studies (50%) that used MAT for diagnosis, often with different antibody titre cut-offs, and the remaining study did not specify the diagnostic method used [Reference Stevens, Carter, Kiep, Stevenson and Schneeweiss50] (see Table 3).
Leptospirosis outcomes, diagnostic tests, serogroups detected, and drivers investigated in included studies
Abbreviations: CI, confidence interval; ELISA, enzyme-linked immunosorbent assay; MAT, microscopic agglutination test; NA, not applicable; NR, not reported; PCR, polymerase chain reaction.
Note: statistically significant drivers of infection (P < 0.05 or if no P-value reported, significance as reported by authors) presented in bold. If unadjusted and adjusted estimates were reported, the driver of infection is bolded if adjusted estimates were significant (refer to Supplementary Material for further details).
Risk of bias within studies
As described previously, 39 studies (93%) were classified as level IV, the lowest level of evidence on the NHMRC evidence hierarchy.
Using the MEVORECH tool, potential biases were identified for all studies across several domains (see Figure 2). The authors did not state the validity and reliability of their methods to measure human Leptospira infection and/or the drivers of infection and often used medical records to sample data rather than methods designed specifically for the purpose of their study. Several studies additionally included probable cases identified from a single positive ELISA and/or MAT result (see Table 3). Furthermore, some studies did not assess the duration and frequency of exposure to drivers of infection, primarily for occupational and lifestyle factors. Lastly, many studies did not account for confounding factors and some failed to provide an effect size with confidence intervals for their estimates.
Traffic light plot representing study-by-study potential biases for each domain of the Methodological Evaluation of Observational Research Checklist (MEVORECH) tool [Reference Shamliyan, Kane, Ansari, Raman, Berkman and Grant12] using the risk-of-bias visualization tool ‘robvis’ [Reference McGuinness53]. Note: 31 studies were considered not applicable for D10 as they did not assess for confounding factors.
Synthesis of results
Sociodemographic drivers
There were 31 studies (31/42, 74%) that reported on sociodemographic drivers of infection including gender/sex (n = 24, 57%), age (n = 15, 36%), ethnicity (n = 7, 17%), poverty (n = 3, 7%), household water supply (n = 3, 7%), awareness of leptospirosis (n = 3, 7%), and education level (n = 1, 2%) (see Supplementary Table S2 for details).
Males were associated with a significantly higher risk of infection compared to females in most studies where gender/sex was investigated (14/24, 58%). One study of army blood bank donors in Hawaii reported higher seroprevalence in females [Reference Lettieri, Moon, Hickey, Gray, Berg and Hospenthal39], and other studies found no association with gender/sex in occupational settings [15, Reference Pittavino, Dreyfus, Heuer, Benschop, Wilson and Collins-Emerson16, 29, Reference Dreyfus, Wilson, Collins-Emerson, Benschop, Moore and Heuer30, 47, Reference Schollum and Blackmore49]. Older individuals, particularly those aged between 20 and 60 years, had a higher infection risk in several studies (6/15, 40%), including in New Zealand [19, Reference Dreyfus, Benschop, Collins-Emerson, Wilson, Baker and Heuer29], New Caledonia [14, Reference Brethes, Puech, Fraisse, Dubois, Domenech and Bourdin25, 34], and Hawaii [Reference Katz, Ansdell, Effler, Middleton and Sasaki33]. However, in Futuna [Reference Massenet, Couteaux and Goarant41] and in a New Caledonia notification data study [Reference Klement-Frutos, Tarantola, Gourinat, Floury and Goarant34], school-aged children and young adults had the greatest risk of infection. Furthermore, in Fiji [Reference Lau, Watson, Lowry, David, Craig and Wynwood38] and New Zealand [Reference Blackmore and Schollum22], there was higher seroprevalence reported in Indigenous populations compared to other ethnicities.
Poverty was another significant driver of infection reported in each of the three studies where it was investigated. The highest seroprevalence was identified in households with an annual income between 20,000 and 30,000 USD in American Samoa [Reference Lau, Dobson, Smythe, Fearnley, Skelly and Clements36] and in Fijian communities with a poverty rate greater than 40% [38, Reference Mayfield, Lowry, Watson, Kama, Nilles and Lau43]. In addition, households with an untreated water supply were associated with a higher risk of infection in Fiji [Reference Lau, Watson, Lowry, David, Craig and Wynwood38], Hawaii [Reference Sasaki, Pang, Minette, Wakida, Fujimoto and Manea48], and Vanuatu (unadjusted estimates only) [Reference Pakoa, Soupe-Gilbert, Girault, Takau, Gaviga and Gourinat44].
Occupational drivers
There were 19 studies (19/42, 45%) that investigated occupational drivers of infection, including meat worker-specific factors (n = 6, 14%), farm worker-specific factors (n = 6, 14%), personal protective equipment (PPE) (n = 6, 14%), time worked within an occupation (n = 5, 12%), type of occupation (n = 4, 10%), indoor-versus-outdoor occupation (n = 3, 7%), animal urine contact (n = 2, 5%), and veterinarian-specific factors (n = 1, 2%) (see Supplementary Table S3 for details).
In New Zealand, leptospirosis incidence rates were highest in meat and farm workers [Reference Thornley, Baker, Weinstein and Maas51]. In meat workers, high-risk work positions were working on the slaughterfloor, offal removal, and stunning/pelting [Reference Pittavino, Dreyfus, Heuer, Benschop, Wilson and Collins-Emerson16, Reference Blackmore and Schollum22, Reference Dreyfus, Benschop, Collins-Emerson, Wilson, Baker and Heuer29, Reference Dreyfus, Wilson, Collins-Emerson, Benschop, Moore and Heuer30]. Two studies (2/3, 67%) reported increased seroprevalence in pig workers compared to workers processing other meats [Reference Blackmore, Bell and Schollum21, Reference Blackmore and Schollum22], whereas another study reported higher annual infection risk in sheep abattoir workers than in beef or deer workers [Reference Dreyfus, Heuer, Wilson, Collins-Emerson, Baker and Benschop15]. In farm workers, farming deer was associated with significantly higher seroprevalence compared to farming beef or sheep, and farmers assisting in calving or fawning also had a higher risk of infection [Reference Sanhueza, Heuer, Wilson, Benschop and Collins-Emerson46]. Moreover, the primary farm-related characteristic associated with higher seroprevalence in workers was the flat terrain of the farm being greater than 25% [Reference Sanhueza, Heuer, Wilson, Benschop and Collins-Emerson46]. Five studies (5/6, 83%) that investigated PPE use in farmers or meat workers reported no reduction in infection risk, irrespective of the type of PPE. In fact, one study instead found that always or often wearing a facemask or safety glasses increased the odds of seropositivity in meat workers [Reference Dreyfus, Benschop, Collins-Emerson, Wilson, Baker and Heuer29].
In American Samoa and Fiji, three studies (3/3, 100%) found that seropositive cases were significantly more likely to be outdoor workers compared to indoor workers [Reference Lau, Clements, Skelly, Dobson, Smythe and Weinstein35, Reference Lau, Dobson, Smythe, Fearnley, Skelly and Clements36, Reference Lau, Watson, Lowry, David, Craig and Wynwood38]. In New Caledonia, one study (1/2, 50%) identified higher seroprevalence in farmers compared to other occupations [Reference Brethes, Puech, Fraisse, Dubois, Domenech and Bourdin25]. Lastly, in the Federated States of Micronesia, those with infection were significantly more likely to be students compared with farmers or other occupations [Reference Colt, Pavlin, Kool, Johnson, McCool and Woodward26].
Lifestyle drivers
There were 15 studies (15/42, 36%) that reported on lifestyle drivers of human Leptospira infection, including water-associated exposures (n = 11, 26%), non-water-based recreational activities (n = 10, 24%), smoking (n = 3, 7%), and open skin wounds (n = 2, 5%) (see Supplementary Table S4 for details).
Swimming was found to be a significant driver of infection in several studies (3/5, 60%), including in New Caledonia [Reference Goarant, Laumond-Barny, Perez, Vernel-Pauillac, Chanteau and Guigon32], American Samoa [Reference Lau, Dobson, Smythe, Fearnley, Skelly and Clements36], and French Polynesia [Reference Coudert, Beau, Berlioz-Arthaud, Melix, Devaud and Boyeau27]. In the Federated States of Micronesia, gardening or tending to crops also increased the risk of infection [Reference Colt, Pavlin, Kool, Johnson, McCool and Woodward26]. In contrast, other recreational activities such as fishing, hunting, and camping were generally not associated with infections [Reference Berlioz-Arthaud, Kiedrzynski, Singh, Yvon, Roualen and Coudert20, Reference Pakoa, Soupe-Gilbert, Girault, Takau, Gaviga and Gourinat44, Reference Sanhueza, Heuer, Wilson, Benschop and Collins-Emerson46]. Furthermore, home slaughter of animals was associated with higher seroprevalence in veterinarians [Reference Sanhueza, Heuer, Wilson, Benschop and Collins-Emerson47] but not in meat or farm workers [Reference Pittavino, Dreyfus, Heuer, Benschop, Wilson and Collins-Emerson16, Reference Dreyfus, Wilson, Collins-Emerson, Benschop, Moore and Heuer30, Reference Sanhueza, Heuer, Wilson, Benschop and Collins-Emerson46].
In addition, one study (1/2, 50%) reported a significant association between skin wounds and Leptospira infection [Reference Sasaki, Pang, Minette, Wakida, Fujimoto and Manea48].
Environmental drivers
There were 29 studies (29/42, 69%) that reported on environmental drivers of human Leptospira infection, including location (n = 16, 38%), animal exposures (n = 13, 31%), rainfall (n = 5, 12%), seasonality (n = 4, 10%), geographical characteristics of households (n = 4, 10%), vegetation/soil type (n = 2, 5%), and human population density (n = 1, 2%) (see Supplementary Table S5 for details).
Geographical locations in the Pacific Islands with significantly higher rates of infection included La Nera in New Caledonia [Reference Brethes, Puech, Fraisse, Dubois, Domenech and Bourdin14, Reference Brethes, Puech, Fraisse, Dubois, Domenech and Bourdin25], Hawaii, and Kauai counties compared to Oahu in Hawaii [Reference Katz, Ansdell, Effler, Middleton and Sasaki33] and the North Island in New Zealand [Reference Blackmore, Bell and Schollum21]. In American Samoa, serogroup Australis was significantly more common in the Manu’a Islands, whereas serogroup Hebdomadis was only identified in Tutuila [Reference Lau, Skelly, Smythe, Craig and Weinstein37]. Furthermore, two studies (2/3, 67%), both conducted in Fiji, identified higher seroprevalence in rural settlements compared to urban residences [Reference Lau, Watson, Lowry, David, Craig and Wynwood38, Reference Mayfield, Lowry, Watson, Kama, Nilles and Lau43].
In terms of animal contact, rodent exposure was a significant factor in several studies (3/6, 50%), including in New Caledonia [Reference Brethes, Puech, Fraisse, Dubois, Domenech and Bourdin14, Reference Goarant, Laumond-Barny, Perez, Vernel-Pauillac, Chanteau and Guigon32] and Fiji [Reference Lau, Watson, Lowry, David, Craig and Wynwood38]. Exposure to pigs and/or cows was also significantly associated with infections in New Caledonia [Reference Goarant, Laumond-Barny, Perez, Vernel-Pauillac, Chanteau and Guigon32], Fiji [Reference Lau, Watson, Lowry, David, Craig and Wynwood38], and Vanuatu (unadjusted estimates only) [Reference Pakoa, Soupe-Gilbert, Girault, Takau, Gaviga and Gourinat44]. Furthermore, local cattle density was strongly correlated with infections in Fiji [Reference Lau, Watson, Lowry, David, Craig and Wynwood38, Reference Mayfield, Lowry, Watson, Kama, Nilles and Lau43] and New Zealand [Reference Thornley, Baker, Weinstein and Maas51].
With respect to seasonality – in New Zealand and New Caledonia, there was a lower incidence of leptospirosis during the cool and dry months compared to the warmer months [Reference Cowie and Bell28, Reference Klement-Frutos, Tarantola, Gourinat, Floury and Goarant34]. In Fiji, infections were significantly associated with floods [Reference Togami, Kama, Goarant, Craig, Lau and Ritter52] and maximum rainfall in the wettest month [Reference Lau, Watson, Lowry, David, Craig and Wynwood38, Reference Mayfield, Lowry, Watson, Kama, Nilles and Lau43]. Leptospirosis cases were also strongly correlated with rainfall observed two months earlier in Futuna [Reference Massenet, Yvon, Couteaux and Goarant42].
Other relevant characteristics of households that were associated with higher seroprevalence included the presence of a water body located less than 100 m from home in Fiji [Reference Lau, Watson, Lowry, David, Craig and Wynwood38, Reference Mayfield, Lowry, Watson, Kama, Nilles and Lau43], house altitude below the median altitude of the village, households located nearby and below piggeries, agricultural vegetation type, and clay loam soil types in American Samoa [Reference Lau, Clements, Skelly, Dobson, Smythe and Weinstein35, Reference Lau, Dobson, Smythe, Fearnley, Skelly and Clements36].
Discussion
This review provides insight into the various drivers of human Leptospira infection in the Pacific Islands. The identified drivers of infection were predominantly occupational in the temperate climate of New Zealand, whereas in the tropical/sub-tropical Pacific Islands, exposures were attributed to a combination of factors including low socioeconomic status, agricultural activity, water-associated exposures, diverse mammalian reservoirs, and weather events. The main drivers of infection identified in this review are shown in Figure 3.
Primary drivers of human Leptospira infection in the Pacific Islands.
Significant drivers of human Leptospira infection in the Pacific Islands
In terms of socio-demographics, males aged between 20 and 60 years were at the highest risk of Leptospira infection in the Pacific Islands, which is consistent with the global literature [Reference Costa, Hagan, Calcagno, Kane, Torgerson and Martinez-Silveira1]. A higher infection risk was identified in Indigenous populations and may be linked to cultural differences in animal husbandry or slaughtering practices [Reference Lau, Watson, Lowry, David, Craig and Wynwood38]. Poverty was another important driver of infection in remote Pacific Islands, which aligns with findings from other resource-limited settings in Asia and South America [Reference Soo, Khan and Siddiqui54]. Low socioeconomic areas are associated with more intimate contact with animals and poor sanitation and hygiene [Reference Lau, Smythe, Craig and Weinstein3]. Furthermore, untreated household water supply systems were independently associated with infections, which may be due to high rates of bacterial contamination in these water sources [Reference Bierque, Thibeaux, Girault, Soupe-Gilbert and Goarant55].
Similar to other temperate regions like Europe [Reference Jansen, Schoneberg, Frank, Alpers, Schneider and Stark56], Leptospira infection in New Zealand was primarily associated with occupational exposures [Reference Thornley, Baker, Weinstein and Maas51]. High-risk abattoir roles were stunning, working on the slaughterfloor and offal removal. Stunning animals can cause splashing of urine, and both slaughterfloor workers and those in offal removal directly handle animal viscera [Reference Dreyfus, Wilson, Collins-Emerson, Benschop, Moore and Heuer30]. Offal removal is also labour-intensive, resulting in skin wounds that may facilitate the transmission of bacteria [Reference Pittavino, Dreyfus, Heuer, Benschop, Wilson and Collins-Emerson16]. Sheep abattoir workers were at the highest risk of infection in recent studies, with workers reporting that sheep tended to urinate spontaneously when stunned [Reference Dreyfus, Benschop, Collins-Emerson, Wilson, Baker and Heuer29]. In older studies, pig workers had an increased infection risk; however, dry stock vaccinations were not available at the time [Reference Blackmore, Bell and Schollum21]. Interestingly, PPE use did not reduce the risk of Leptospira infection in workers and one study reported an increased risk of infection with facemask and safety goggles’ use [Reference Dreyfus, Benschop, Collins-Emerson, Wilson, Baker and Heuer29]. Whilst the finding of an increased infection risk with PPE use was obtained from unadjusted estimates and thus likely influenced by confounding factors [Reference Dreyfus, Benschop, Collins-Emerson, Wilson, Baker and Heuer29], the lack of reduction in infection risk with PPE use across studies warrants further investigation. A possible reason may be poor compliance, as workers stated that they often lifted up their safety goggles or masks to remove sweat or fog to restore visibility [Reference Dreyfus, Benschop, Collins-Emerson, Wilson, Baker and Heuer29]. These findings warrant further investigation into the use of PPE in abattoir workers. In farm workers, farming deer was associated with a higher risk of infection compared to farming other animals [Reference Sanhueza, Heuer, Wilson, Benschop and Collins-Emerson46]. Deer are not used to being handled by farmers, and they were reported to dribble urine whilst being handled [Reference Sanhueza, Heuer, Wilson, Benschop and Collins-Emerson46]. Additional factors associated with Leptospira infection in farm workers were assisting in cattle and deer calving, which would place workers into close contact with animal urine and placental tissue, and having a farm with flat terrain, which may increase the presence of standing puddles that accumulate leptospires [Reference Sanhueza, Heuer, Wilson, Benschop and Collins-Emerson46]. For veterinarians, home slaughter of animals was a significant driver of infection [Reference Sanhueza, Heuer, Wilson, Benschop and Collins-Emerson47], which could be linked to a lack of familiarity with appropriate hygiene practices for animal slaughter.
In tropical Pacific Islands, farmers were an at-risk occupational group [Reference Brethes, Puech, Fraisse, Dubois, Domenech and Bourdin25], and accordingly, rural agricultural areas were geographical hotspots for infection [Reference Brethes, Puech, Fraisse, Dubois, Domenech and Bourdin14, Reference Brethes, Puech, Fraisse, Dubois, Domenech and Bourdin25, Reference Katz, Ansdell, Effler, Middleton and Sasaki33, Reference Lau, Watson, Lowry, David, Craig and Wynwood38, Reference Mayfield, Lowry, Watson, Kama, Nilles and Lau43]. Students were another high-risk group, which is likely attributed to their recreational exposures [Reference Colt, Pavlin, Kool, Johnson, McCool and Woodward26, Reference Klement-Frutos, Tarantola, Gourinat, Floury and Goarant34]. Freshwater swimming was a significant lifestyle driver of infection across tropical Pacific Islands and has been recognized as a high-risk activity for leptospirosis in the literature due to the ability of leptospires to remain virulent in freshwater for several weeks [Reference Bierque, Thibeaux, Girault, Soupe-Gilbert and Goarant55]. Furthermore, gardening was associated with infections in the Federated States of Micronesia [Reference Colt, Pavlin, Kool, Johnson, McCool and Woodward26], which brings individuals into close contact with rodents and soil and water contaminated with their urine. In addition, open skin wounds may predispose individuals to infection, facilitating the transmission of leptospires to humans by compromising the physical dermal barrier [Reference Gostic, Wunder, Bisht, Hamond, Julian and Ko57].
The predominant animal reservoirs for infection in tropical Pacific Islands were rodents and livestock (especially cattle and pigs). Traditionally, livestock are recognized as important hosts of leptospirosis in rural areas and rodents are the main reservoirs in urban areas [Reference Soo, Khan and Siddiqui54]. The remote Pacific Islands represent a unique situation with frequent exposure to both rodents and livestock due to the integration of agriculture into urban residences, for example, through backyard piggeries [Reference Lau, Dobson, Smythe, Fearnley, Skelly and Clements36] and household crops [Reference Guernier, Goarant, Benschop and Lau5].
In terms of seasonality, higher leptospirosis incidence occurred during the wet season in tropical Pacific Islands [Reference Berlioz-Arthaud, Kiedrzynski, Singh, Yvon, Roualen and Coudert20]. Similar findings have been reported in other tropical locations, including Far North Queensland in Australia [Reference Fairhead, Smith, Sim, Stewart, Stewart and Binotto58] and Sri Lanka [Reference Robertson, Nelson and Stephen59]. During the wet season, there are dissemination of leptospires from rainfall [Reference Kupek, de Sousa Santos Faversani and de Souza Philippi60], an increase in recreational water-associated exposures, and the proliferation of rodent populations during their breeding season [Reference Brown, Aplin, Hinds, Jacob, Thomas and Ritchie61]. In the Pacific Islands, the highest risk of infection was in the one-to-two-month period following heavy rainfall events [24, Reference Massenet, Yvon, Couteaux and Goarant42]. Short periods of intense rainfall result in the accumulation of leptospires in soil and water at the surface level [Reference Bierque, Thibeaux, Girault, Soupe-Gilbert and Goarant55] whilst simultaneously contaminating water supply systems and displacing people from their homes and animals from their habitats, transiently increasing human–animal interactions [Reference Lau, Smythe, Craig and Weinstein3].
The abundance of backyard piggeries in the remote Pacific islands additionally poses a distinct risk to these communities. Piggeries are typically built adjacent to rivers which allow drainage of waste but also result in contaminated water travelling downstream towards other households [Reference Lau, Watson, Lowry, David, Craig and Wynwood38].
Implications for mitigation strategies
To target the main drivers of infection identified in this review, mitigation strategies can be implemented at the regional, community, and individual levels.
At a regional level, improving the disease reporting would enhance our understanding of leptospirosis epidemiology. MAT and PCR diagnosis require the transport of specimens to laboratories with specialized equipment [Reference Niloofa, Fernando, de Silva, Karunanayake, Wickramasinghe and Dikmadugoda62], which is often not feasible in remote Pacific Islands. Hence, the use of more readily available tests such as IgM ELISA and IgM rapid immunochromatography test (e.g. Leptocheck-WB) may help to improve disease reporting and response to outbreaks. These tests have demonstrated high diagnostic sensitivity during acute leptospirosis illness [Reference Niloofa, Fernando, de Silva, Karunanayake, Wickramasinghe and Dikmadugoda62] and are now being used for initial diagnosis in French Polynesia [Reference Bouscaren, de Coignac, Lastere, Musso, Teissier and Formont24], Wallis and Futuna [Reference Massenet, Couteaux and Goarant41], Vanuatu [Reference Pakoa, Soupe-Gilbert, Girault, Takau, Gaviga and Gourinat44], and Fiji [Reference Togami, Kama, Goarant, Craig, Lau and Ritter52]. In addition, greater collaboration with veterinary surveillance systems for leptospirosis could help to inform human infection risk and public health response [Reference Goarant63].
Furthermore, increasing the use of animal vaccinations against leptospirosis could further reduce disease prevalence in animal reservoirs, particularly in New Zealand. Whilst there are compulsory vaccination programmes for dairy cattle and pigs, there remains limited uptake of the available vaccines for sheep and deer [Reference Yupiana, Wilson, Collins-Emerson, Weston, Benschop and Vallee64]. Given the risk of infection associated with both sheep and deer handling, promoting the uptake of these vaccinations could further reduce transmission risk for New Zealand workers.
At a community level, improving water security, pest control, and piggery management could significantly reduce the leptospirosis burden in remote Pacific Islands. Climate-resilient water sanitation projects have been implemented with success in the Solomon Islands [65] and should be promoted in other islands to reduce the risk of household water supply contamination. Local communities can also contribute by maintaining sewage systems and preventing the accumulation of debris in waterways [Reference Lau, Smythe, Craig and Weinstein3]. Improved waste management would assist in controlling rodent populations, and additional measures such as rat trapping, removing nesting places, and limiting rodent access to food could be implemented during rodent breeding months [Reference Brown, Aplin, Hinds, Jacob, Thomas and Ritchie61]. Furthermore, proper management of piggery waste and aiming to build piggeries further away and downhill from homes could mitigate the risk for households located nearby and below piggeries [Reference Lau, Clements, Skelly, Dobson, Smythe and Weinstein35].
Further adjustments to occupational practices in New Zealand and targeted information campaigns in remote Pacific Islands could address the individual-level drivers of infection. In abattoirs, employers could introduce regular PPE compliance audits and institute additional measures such as fit testing for facemasks and safety goggles. To reduce the infection risk for slaughterboard workers, stunned animals could be placed on a separate platform [Reference Dreyfus, Benschop, Collins-Emerson, Wilson, Baker and Heuer29], or on disposable covers, to minimize the accumulation of urine. For workers in offal removal, the use of cut-resistant gloves could decrease the risk of lacerations to workers’ hands [Reference Caple66]. In remote Pacific Islands, public health media campaigns could promote messages about simple hygiene risks (e.g. ‘wash wounds with clean water and soap, and cover up’ and ‘use boots and gloves when gardening or in the field’), reducing rodent infiltration (‘avoid leaving rubbish scattered around the home’) [Reference Goarant63] and minimizing freshwater exposures (‘avoid swimming in puddles and rivers after heavy rain’). Furthermore, in New Caledonia and French Polynesia, leptospirosis education has been incorporated into the primary school curriculum through posters and board games [Reference Goarant63], and other Pacific Islands could utilize similar approaches to promote preventative behaviours from a young age. In addition, collaboration with Indigenous community leaders is required to identify and address cultural activities that may contribute to a higher infection risk.
Limitations
This review provides important insight into the drivers of human Leptospira infection in the Pacific Islands; however, there were some limitations present. Given the broad nature of this review, there was significant heterogeneity of included studies, and hence, meta-analysis was not conducted, and instead, study findings were synthesized descriptively. Furthermore, the MEVORECH tool is validated for risk factor studies but was originally designed for chronic diseases [Reference Shamliyan, Kane, Ansari, Raman, Berkman and Grant12], and hence, potential biases identified in domains such as the reference period for outcomes and/or exposures may be attributed to the limited applicability of the tool to acute infectious disease studies. Lastly, several studies in this review included probable cases diagnosed using indirect methods, such as ELISA and MAT, often with differing antibody titre cut-off levels, and sometimes only reported unadjusted estimates, which made it challenging at times to identify the true drivers of infection.
Directions for future research
Future research on human leptospirosis in the Pacific Islands designed to minimize bias is essential to further our knowledge of local disease epidemiology. There is a need for greater focus on serogroup-specific drivers of Leptospira infection and studies at a regional level that compare findings in each Pacific Island to individualize mitigation strategies. Furthermore, research into PPE use in meat workers is required given the limited effectiveness identified in this review.
It is also important to consider how climate change may exacerbate existing drivers of leptospirosis in the region. The Pacific represents one of the most natural disaster-prone regions in the world [65], and the predicted increase in floods and cyclones may increase opportunities for Leptospira infection by disrupting local ecosystems, contaminating water sanitation networks, and damaging human infrastructure [Reference Lau, Smythe, Craig and Weinstein3]. We advocate for ecosystem-based adaptation approaches, such as mangrove conservation and restoration projects, which can provide coastal flood protection and useful economic and environmental co-benefits for Pacific Island communities [Reference Prasad, Waqa-Sakiti and Viliamu Iese67, Reference Menéndez, Losada, Torres-Ortega, Narayan and Beck68]. With a significant number of leptospirosis outbreaks after flooding being reported worldwide, the mitigation strategies developed by the Pacific islands may help to inform the actions of other affected regions.
Conclusions
The drivers of human Leptospira infection in the Pacific Islands are multifactorial and differ significantly between New Zealand and the tropical Pacific Islands. The drivers of infection in New Zealand exemplify the predominance of occupational risks in temperate climates, and public health interventions should accordingly target abattoir and farm workers. In tropical Pacific Islands, infections are the product of distinct environmental characteristics, diverse animal reservoirs, and human activity that facilitate leptospirosis transmission. The drivers of infection identified in this review are not necessarily distinct from factors identified in other parts of the world; however, it is the complexity and multifaceted nature of leptospirosis transmission that is unique to the region, particularly in the face of climate change. Leptospirosis is ultimately a preventable disease, but further research is required to untangle the ecological mechanisms that underlie disease transmission in the Pacific Islands.
Supplementary material
The Supplementary Material for this article can be found at http://doi.org/10.1017/S0950268824001250.
Data availability statement
All relevant data are provided in text and in the Supplementary Material.
Acknowledgements
We would like to thank those in the School of Public Health at the University of Adelaide who provided feedback for the initial manuscript.
Author contribution
All authors conceptualized the study, designed the methodology, and assisted in study selection, data extraction and formal analysis. S.K. wrote the original draft. All authors contributed to drafting and revising the manuscript.
Funding statement
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Competing interest
The authors declare none.
Open access
