ENVIRONMENTAL FACTOR INVESTIGATION OF EXUDATIVE CLOACITIS IN KĀKĀPŌ (STRIGOPS HABROPTILUS) ON WHENUA HOU (CODFISH ISLAND), NEW ZEALAND

The Kākāpō (Strigops habroptilus) is a critically endangered, nocturnal parrot endemic to New Zealand with an estimated total population of 201 adult birds in the wild (DOC 2021). The Kākāpō population has declined primarily since European settlement of New Zealand, due to predation, habitat loss, and habitat modification (Bergner et al. 2016). During the 1970s, intensive research and conservation management of kakapo was started from two remnant populations (Clout and Merton 1998; Elliott et al. 2001).

As a result of intensive management, the Kākāpō population has increased from 51 birds in 1995 (Powlesland et al. 2006) to its current size, and breeding populations are managed on three predator-free offshore islands: Whenua Hou (Codfish Island), Te Hauturu-o-Toi (Little Barrier Island), and Pukenui (Anchor Island), New Zealand (Fig. 1). These three populations currently require intensive management (Clout 2006; Jansen 2006) and regular supplementary feeding.

Since 2002, a disease called exudative cloacitis has emerged in the Kākāpō on Whenua Hou (Jakob-Hoff and Gartrell 2012). In Kākāpō, exudative cloacitis refers to an inflammation of the vent margin or cloaca, which is often ulcerated and covered in crusty exudate (Jakob-Hoff and Gartrell 2012). The condition is debilitating and can lead to reduced fertility.

The factors responsible for initiating exudative cloacitis in Kākāpō are unknown despite extensive pathological and microbiological investigations (Jakob-Hoff et al. 2009; Jakob-Hoff and Gartrell 2012). The disease is geographically limited to Whenua Hou, supporting the hypothesis that the etiology of the exudative cloacitis may involve one or more as yet unidentified environmental factors in the Kākāpō habitat. A similar condition seen in poultry, vent gleet, is associated with irritation of the vent by old, heavily fecally contaminated litter or diarrhea. It is suggested that this condition in poultry is often not involved with any primary pathogen but is instead associated with a combination of infectious and noninfectious factors (Crosta et al. 2003).

It is possible that the exudative cloacitis in Kākāpō is associated with some island-specific factors leading to irritation, inflammation, and ulceration of the mucocutaneous border of the cloaca. The initiating causes of the exudative cloacitis may be masked by secondary infection and inflammation associated with the contact of feces, urine, and urates on the ulcerated mucosa. Identification of possible contributing factors is critical in deciding on treatment and prevention of this condition in Kākāpō.

The description of disease frequency in individuals, and in time and space in different populations, is a key method used in epidemiology to investigate a disease outbreak (Wobeser 2007). The aim of our research was to perform an epidemiological study to investigate the environmental factors contributing to the initiation of exudative cloacitis in Kākāpō by 1) producing and describing a case series, 2) mapping the geographic distribution of exudative cloacitis cases, 3) investigating the chemical characteristics of Kākāpō roosting sites, and 4) assessing the effects of climatic factors on the incidence of exudative cloacitis each year. This epidemiologic investigation was conducted as part of combined research to investigate the pathogenesis of exudative cloacitis in Kākāpō; the clinicopathological, microbiological, and genomic investigations are beyond the limits of this study.

Kākāpō health data were obtained from a surveillance program managed by the Kākāpō Recovery Team, Department of Conservation (DOC), New Zealand. This contained all the data collected during annual transmitter changes, health checks, and any other opportunistic observations made on individual Kākāpō during field visits. Usually, health examinations of Kākāpō were performed once a year; individuals with a history of exudative cloacitis have been examined four times per year in many years. In 2017, all the Kākāpō were examined four times for possible signs of exudative cloacitis. Variables recorded were the identification of the Kākāpō, sex, age, date of observation, location data as easting and northing (New Zealand Transverse Mercator), and exudative cloacitis status at the time of observation. The exudative cloacitis case definition used for the surveillance data was the presence of inflammation, ulcerations of the mucocutaneous margins of the cloaca, or accumulation of exudates around and within the cloaca (Jakob-Hoff and Gartrell 2012). A grading scheme to categorize exudative cloacitis in Kākāpō has been developed by the DOC. Depending on the severity of the lesion on the cloaca or the vent, cases have been graded as mild, moderate, and severe (Fig. 2). All new Kākāpō cases (incident cases) diagnosed with exudative cloacitis in 2002–17 were selected for the case series. Descriptive statistical analysis was done using R version 3.4.1 (R Studio Team 2016) to identify the trend of disease occurrence over the years and to describe the birds diagnosed with the disease by age, sex, location, and if they belonged to the founder population or not.

The age descriptor for older Kākāpō of unknown age was calculated assuming a minimum of 10 yr old on the date of their first discovery. Age groups as chicks (0–149 d), juveniles (150 d to 5 yr), and adults (>5 yr) were determined by considering the average age of fledging and sexual maturity of Kākāpō on offshore islands (Eason et al. 2006; Farrimond et al. 2006b).

The distribution of exudative cloacitis cases was mapped for the period 2002–17 on Whenua Hou and used to describe the pattern of the disease. Point maps were developed in R version 3.4.1 (R Studio Team 2016) and Quantum GIS version 3.8.1 (QGIS Development Team 2018) using the geographic locations of the individual Kākāpō on Whenua Hou each year for diseased and healthy birds. Point maps were created using packages rgdal (Bivand et al. 2019a), ggmap (Kahle and Wickham 2013), ggplot2 (Wickham 2016), epiR (Stevenson et al. 2013), ClassInt (Bivand et al. 2007), RcolorBrewer (Neuwirth 2014), plyr (Wickham 2011), rgeos (Bivand et al. 2019c), sp (Bivand et al. 2013), maptools (Bivand et al. 2019b), raster (Hijmans and van Etten 2012), spatstat (Baddeley et al. 2015), and scales (Wickham 2017) for R.

To analyze the spatial point pattern and identify a possible spatial clustering of cases, the Kfunction was calculated for both case and healthy birds. The difference between these two Kfunctions was examined as it provides an effective summary of spatial dependence over a large distance (Diggle et al. 1995; Carpenter 2001; Böhm et al. 2008; Mountrakis and Gunson 2009). The K-function of a point pattern is defined as the expected number of farther points within a distance h of an arbitrary point, divided by the overall density of points (Ripley 1976). The Kfunction analysis was performed using packages fields (Nychka et al. 2017), SpatialEpi (Chen et al. 2018), and gridExtra (Auguie and Antonov 2017) for R. To provide 95% confidence intervals (CIs) for this function, Monte Carlo simulations of the Poisson point pattern process were used (Harrison 2010).

To investigate the environmental characteristics of the Kākāpō roosting sites on Whenua Hou, three chemical parameters of the topsoil layer of the roosting sites were examined. The same chemical parameters were analyzed from the topsoil layer of the general areas where Kākāpō roosts were not found on Whenua Hou and on Te Hauturu-o-Toi, as control samples. Sample collections were conducted from August 2017 to July 2018 on Whenua Hou and during October 2018 on Te Hauturu-o-Toi. Samples were transported in polyethylene bags and stored at 4 C until analyzed to determine pH, ammonium, and moisture contents.

On Whenua Hou, 49 soil samples were collected from the Kākāpō roosting sites, including six samples from the roosts of diseased Kākāpō at the time of collection. From the general areas of the Whenua Hou and Te Hauturu-o-Toi, 27 and 32 soil samples were collected respectively. Moisture content of the soil samples was measured gravimetrically, after oven drying a known amount of sample to a constant weight for 12 h at 105 C and calculating the weight loss of the sample (Stumpy and Binkley 1993). The pH of the roost materials was measured in deionized water using a pH meter (HI 2211 Microprocessor-based pH/mV/°C Bench Meters, Hanna Instruments, Nusfalau, Romania). For each soil sample, 10 g was mixed with 25 mL of deionized water; the pH of the solution was measured after 12 h. Exchangeable ammonium (NH₄⁺) that can be extracted at room temperature with a neutral potassium salt solution was measured. Ammonium was extracted by shaking with 2.0 M potassium chloride as described (Keeney and Nelson 1982; Maynard and Kalra 1993). The amount of ammonium in the extract was determined by colorimetric techniques (Maynard and Kalra 1993). Box plots were made to visualize the distribution of moisture content, pH, and ammonium in Kākāpō roost and general areas of Whenua Hou and general areas of Whenua Hou and Te Hauturu-o-Toi using the R packages ggplot2 (Wickham 2016) and EnvStats (Millard 2013). The mean and CIs for the pH, moisture content, and ammonia were calculated as a summary statistic in R version 3.4.1 (R Studio Team 2016). Soil moisture content, ammonia, and pH comparisons were made by the Welch t-test in R version 3.4.1.

Climate data for Whenua Hou could not be collected for 2002–17 because the island had no functioning weather station for that period. Therefore, the data recorded from the nearest located weather station on Rakiura (Stewart Island; Fig. 1) was used. Available data for wind direction, wind speed, wind run, relative humidity, rainfall, dry temperature, wet temperature, and annual maximum and minimum temperatures of Rakiura (1987–93 and 2002–17) and for Whenua Hou (1987–93) were obtained from New Zealand's National Climate Database (NIWA 2018). Before extracting the relevant weather data for Rakiura, correlation between the available weather data of Whenua Hou 1987–93 with that of Rakiura was examined through scatterplots and a Pearson correlation.

Correlation between the data from Rakiura and available data from Whenua Hou were analyzed using library readr (Wickham et al. 2018) in R. To examine the effect of climate on the number of exudative cloacitis cases observed each year, a negative binomial regression analysis was performed in R version 3.4.1 using the MASS package (Venables and Ripley 2002). The number of exudative cloacitis cases per year was used as the outcome variable and the population at risk as an offset. Climate variables that had a Pearson correlation of 0.85 or more between measurements on Rakiura and Whenua Hou were retained and their Rakiura annual average used as independent variables in the model. Incidence risk ratio was calculated by exponentiating the coefficient and the boundaries of the CI.

There were differences in the number of Kākāpō that had undergone a health examination in Whenua Hou in each year (Fig. 3). Exudative cloacitis was diagnosed in 22 Kākāpō on Whenua Hou over the period 2002–17, including 12 males and 10 females. The affected birds belonged to all age groups: 13 adults, seven juveniles, and two chicks. Nine of 22 cases were from the older founder birds that had established the population of Kākāpō on Whenua Hou. Occurrence of slow progressing sporadic cases of different severity were observed over time, with an increased number of cases detected during the years 2015–16. No cases were detected in some years during the study period (Figs. 3 and 4).

Kākāpō with exudative cloacitis were distributed throughout Whenua Hou (Fig. 5) and the disease was recorded in all areas of the island where the Kākāpō population is distributed. The observed K patterns (Fig. 6) of the distribution of diseased and healthy Kākāpō were compared with the theoretical K curve under complete spatial randomness. According to these two graphs, both the diseased and the healthy population of Kākāpō are clustered within all distances. Kākāpō with exudative cloacitis are less clustered, or more spread out, than the healthy Kākāpō population (Fig. 7).

All three chemical parameters examined showed a range of distribution in Kākāpō roosts and general areas of the two islands (Figs. 8 and 9). The mean ammonium content of the few Kākāpō roost samples of diseased birds (138.98 parts per million [ppm]; 95% CI 163.22–441.19 ppm) was higher than the mean ammonium content of the roost samples without cases on Whenua Hou (32.38 ppm, CI 6.83–57.92 ppm). This difference was due to the higher ammonium contents of two samples from the cases. However, this difference was not statistically significant (P=0.384). Statistical comparisons for moisture and pH were not performed because of the low sample size.

The mean soil pH of the roost samples (pH 5.29, CI 5.00–5.58) was lower (P=0.026) (i.e., more acidic) compared with the general areas (pH 5.92, CI 5.48–6.36). The mean moisture content of the roost sites (64.23%, CI 59.53–68.93%) was significantly lower than that of the general areas (P=0.031). However, mean ammonium levels of the roost (32.38 ppm, CI 6.83–57.92 ppm) and general areas (23.43 ppm, CI 9.66–37.20 ppm) of Whenua Hou were not significantly different (P=0.244). The soil analysis from Te Hauturu-o-Toi's general areas, where exudative cloacitis has not yet occurred, showed very strong evidence (P=0.0001) that the pH was more acidic and moisture content was lower (P=0.0014), although there was weak evidence (P=0.041) that the ammonium content was lower in the substrate samples from Te Hauturu-o-Toi in comparison with the general samples from Whenua Hou.

There was a strong positive correlation between the annual maximum temperature and rainfall for Whenua Hou (between 1987 and 1993) and those of Rakiura, with correlation coefficients of 0.88 and 0.86, respectively (Supplementary Material Figs. S1, S2). Therefore, weather data for Rakiura for the period of 2002–17 was used to investigate the effect of annual maximum temperature and rainfall for Whenua Hou. The incidence risk ratios for annual maximum temperature and rainfall were 1.14 and 1.00 respectively, indicating that neither temperature nor rainfall was significantly associated with the number of exudative cloacitis cases in Kākāpō on Whenua Hou each year (Table 1).

Our study confirms that exudative cloacitis in Kākāpō between 2002 and 2017 was geographically restricted to Whenua Hou, suggest an initiating cause for exudative cloacitis that is constrained to Whenua Hou, but within the island there was no evidence of spatial clustering, with disease occurrence mirroring Kākāpō distribution on the island. Although there was evidence of temporal variation in case occurrence between years, with some years having no new cases identified, there was no evidence of temporal clustering of the new cases. The progression of exudative cloacitis has been slow over the years and the pattern appears to be sporadic. The small increase in the number of cases detected in 2015 and 2016 could be either a true increase in the incidence or represent observational bias in rangers detecting more cases as they increased the frequency of clinical examinations per year and had increased awareness about the disease at this point. There was also no evidence for an age or sex predilection: both sexes and a mixed group of adults and juveniles were affected. These epidemiologic descriptors do not support the hypothesis that exudative cloacitis is a transmissible infectious disease, nor do they support the hypothesis of a geographically constrained environmental hazard within Whenua Hou that is affecting the Kākāpō.

In general, Kākāpō are solitary and live in individual home ranges that usually overlap during most of the year (Higgins 1999; Powlesland et al. 2006). Kākāpō home-range sizes were found to be different among their offshore islands around New Zealand (Moorhouse and Powlesland 1991; Powlesland et al. 2006). On Whenua Hou with a dense population of Kākāpō, the mean home-range size recorded for adult females and newly fledged juveniles was approximately 15 ha (Farrimond et al. 2006a). On Whenua Hou, habitat selection of breeding and nonbreeding adult female Kākāpō was found to be dependent on the distribution patterns of different types of vegetation. Research on habitat selection by Kākāpō has shown that they are not distributed randomly across the landscape, but rather prefer some habitats more commonly than others (Moorhouse 1985; Walsh et al. 2006). Therefore, the distribution pattern of birds with exudative cloacitis on Whenua Hou could represent the pattern of their preferred habitats on the island.

Kākāpō usually roost in a shady area on the ground or in natural cavities such as tree logs within their home range. Some roost sites have been used repeatedly or irregularly over many weeks or even years by some individuals (Higgins 1999; Powlesland et al. 2006). All the chemical parameters of the roost sites and other general areas of the two islands examined showed high variability in their distribution. In general, the soil was mostly acidic in all the samples studied and there was a difference between the pH levels detected in roost and general areas of Whenua Hou. Low pH levels have been detected in New Zealand offshore islands with dense populations of seabirds (Mulder and Keall 2001). Nitrification of ammonium in bird feces is one of the mechanisms suspected to be contributing to the soil acidity (Mulder and Keall 2001; Fukami et al. 2006; Otero et al. 2018). Both Whenua Hou and Te Hauturu-o-Toi support large breeding colonies of several species of seabirds and the acidic pH on both islands could be a result of the guano of these birds. Within the Kākāpō roost sites, the higher acidity of the substrate is likely to represent fecal and urate contamination of the roost by the Kākāpō rather than from seabirds, although some roost sharing does occur (Powlesland et al. 2006). Exudative cloacitis has not yet been diagnosed in the currently largest population of Kākāpō, which inhabits Pukenui. Information about the presence of seabird colonies on the island is very limited, although there are some reports mentioning colonies of seabirds in the area (Miskelly 2017). Substrate characteristics such as pH and ammonium of this island and the Kākāpō roost sites there should be investigated to see if the lack of seabirds might be one possible explanation for the absence of the disease on this island.

On Whenua Hou, there was no difference between the mean ammonium content in Kākāpō roost compared with that of the general areas. Uric acid and small amounts of ammonia in bird excreta is one of the main ways that ammonia is added to the soil. The uric acid is mineralized readily to ammonium and available in soil for the terrestrial plants. Volatilization is estimated to release a large amount of this soil ammonium to the atmosphere (Riddick et al. 2012, 2014). It is not clear if the high ammonium content in the roost could have initiated the initial irritation around the cloacal mucosa as occurs in poultry, or whether the ammonium level of the roost soil is elevated because the diseased birds were using the same roost repeatedly. For example, the increased ammonia seen in the roost sites may be due to a difference in diet between affected and unaffected birds, resulting in increased nitrogenous wastes being excreted by diseased birds.

Usually, poultry are exposed to 50 ppm of ammonia in a poultry house atmosphere, but in poorly ventilated housing this level can go as high as 200 ppm, which affects the health status of poultry (Carlile 1984; Ferguson et al.1998). High levels of ammonia in poultry houses are recognized as a cause for keratoconjunctivitis, respiratory infections caused by mucosal damage, and blisters on the breast (Carlile 1984; Ferguson et al. 1998). In our study, ammonia levels in the soil of Kākāpō roosts and general areas of the Whenua Hou and Te Hauturu-o-Toi were lower than the level of ammonia that poultry are usually exposed to in poultry houses. However, the mean ammonia levels of the Kākāpō roosts with diseased birds are higher than the critical level of ammonia in poultry houses. This might suggest ammonia levels in Kākāpō roosts as a predisposing factor for the initiation of exudative cloacitis. There were also significant differences in all three measured chemical properties of the substrates from general areas of Whenua Hou versus Te Hauturu-o-Toi; however, it is not clear if any of these three environmental chemical properties of the soil are implicated in the presence of exudative cloacitis in birds on Whenua Hou but not on Te Hauturu-o-Toi.

If an environmental factor is important in the initiation of exudative cloacitis in Kākāpō, the clinical signs of disease would most probably start as a skin irritation followed by ulceration and inflammation with subsequent secondary infections. This is difficult to confirm because of the necessarily sporadic monitoring of the birds in their island habitats. It has been believed that a combination of wet litter, high ammonia content, and other chemical factors are responsible for the vent and leg dermatitis of broiler chickens (Wang et al. 1998). Experimental studies on foot pad dermatitis in broiler chickens have demonstrated that only the moisture content of the litter has an effect on disease incidence, not pH or ammonia (Martland 1984; Wang et al. 1998). The moisture content of the Kākāpō roosts varied between 24% and 80% in our study, and the mean moisture content was significantly lower than that of the general areas of the island. There was no clear evidence for the effects of moisture or ammonium on the initiation of exudative cloacitis in Kākāpō, but both the ammonia and moisture content of the Kākāpō roosts are important chemical factors to prospectively evaluate with the occurrence of disease in the future.

Temperature, rainfall, and other climatic factors may affect disease incidence, distribution, and severity (Harvell et al. 2002; Rosenthal 2009; Gallana et al. 2013). For Kākāpō, temperature and annual rainfall might affect the availability of food resources, which are important for reproductive success (Elliott et al. 2001), growth of the birds, and development of immunity against diseases, but we were not able to demonstrate any evidence for an effect of the annual maximum temperature and rainfall on the yearly incidence of exudative cloacitis. Weather data were, from necessity, obtained from the neighboring Rakiura. We recommend the reinstatement of weather stations on all of the islands used for Kākāpō conservation to enable assessment of the effects of changing weather patterns in the local climate on this vulnerable species.

Kākāpō are one of the most intensively monitored free-living birds in the world. Individual birds on their island locations are monitored by telemetry and data are collected and recorded accordingly. Nevertheless, there are gaps in the data that limit a more complete investigation of the epidemiology of exudative cloacitis. Our study was conducted retrospectively and the initial data collection for Kākāpō management was not focused on disease surveillance or investigation. In wildlife populations, basic demographic characteristics are often unknown (Tompkins et al. 2015). The gaps in the data include Kākāpō home ranges, detailed environmental characteristics of roosting sites, information on Kākāpō diet, contact with other species including use of burrows and their guano, climatic data of the island over the years, other habitat characteristics (e.g., soil types, vegetation types), early nesting environment, parentage of the founders, and individual health parameters. These gaps may have limited our ability to identify the key risk factors for the initiation of the disease. Identification of these key risk factors would allow more targeted pathologic, environmental, and microbiologic investigations about exudative cloacitis. Further investigation of environmental soil and roost characteristics, such as the ammonia and moisture content of Kākāpō roosting sites, is recommended in future studies when new cases are diagnosed.

The authors thank the members of Kākāpō Recovery Programme, DOC, New Zealand for providing access to Kākāpō data collected over the years: Ian Furkert, Senior Technician and Lance Currie, Senior Technical Manager from the School of Agriculture and Environment for their immense support with Kākāpō roost sample analysis.

Supplementary material for this article is online at http://dx.doi.org/10.7589/JWD-D-21-00201.