The burden of yersiniosis in New Zealand, 2022

Priority setting for infectious disease management at a national level requires risks to be expressed in a consistent manner. Metrics of disease burden have gained wide acceptance for this purpose. These metrics also facilitate the setting of targets for reductions in infectious disease incidence and the assessment of the effectiveness of specific interventions designed to reduce disease.

To measure the burden of disease any chosen metric should integrate the amount (incidence) of a particular disease and its impact (severity) and combine fatal and non-fatal outcomes. Led by the global burden of disease estimates generated by the World Health Organization (WHO),^1,2 the disability-adjusted life year (DALY) metric is gaining increasing acceptance as a measure of the burden of illness. The fundamental calculation for DALYs is: DALY=YLL+YLD (1), where YLL is the number of years of life lost due to mortality and YLD is the number of years lived with a disability, weighted with a factor between zero and one for the severity of the disability. Disability weights are a measure of health state preferences and were originally derived using a personal trade-off (PTO) approach.³

Several studies have used the DALY metric to estimate the burden of diseases, including yersiniosis, in New Zealand.^4–6 However, these studies have often been dependent on characteristics of diseases in other countries to assess the disease burden in New Zealand and, in many cases, it is uncertain whether these attributes are applicable to the diseases in New Zealand.

Yersiniosis is a mainly enteric disease caused primarily by infection with Yersinia enterocolitica (Y. enterocolitica) and, to a lesser extent Y. pseudotuberculosis. Pigs are a recognised reservoir for Y. enterocolitica, and transmission is believed to be predominantly through food.⁷ The disease is diagnosed through the detection of the organism or its nucleic acids in faeces or blood.⁸ The condition is usually self-limiting. While most cases of yersiniosis involve uncomplicated acute gastrointestinal illness, more severe outcomes have been reported for some cases, including colitis, ileitis and pseudoappendicitis.⁹ Reactive arthritis¹⁰ and erythema nodosum¹¹ are recognised sequelae occurring subsequent to a proportion of cases of yersiniosis.

The reported annual incidence of yersiniosis in New Zealand was reasonably stable since the condition first became notifiable in 1996, until 2013, with 330 to 546 cases notified each year.¹² However, since then the incidence has risen steadily to 1,294 in 2022 (25.3 per 100,000).¹³ New Zealand now has the highest reported rate of yersiniosis in the world, with the next highest reported rates of yersiniosis being less than 10 per 100,000.¹³

A 17-month yersiniosis case-control study was conducted in two regions of New Zealand.¹⁴ Cases were administered a questionnaire that allowed quantification of aspects of the burden of disease, in addition to information on risk factors. Data from this study, as well as data from the only New Zealand study on the incidence of acute gastrointestinal illness,¹⁵ now allows the estimation of the burden of yersiniosis in New Zealand for the first time, based largely on New Zealand-specific information.

Methods

Illness outcomes included

Four outcomes for primary illness were defined in terms of acute gastrointestinal illness (AGI): AGI (do not visit a general practitioner [GP] and recover), AGI (visit a GP and recover), AGI (hospitalised and recover) and AGI (death). Hospitalised cases are considered to be a subset of cases visiting a GP, while fatal cases are considered to be a subset of hospitalised cases. Final case numbers in each category were adjusted to avoid double counting. Sequelae following AGI were defined as: reactive arthritis (ReA, subcategories of GP visit and hospitalised) and erythema nodosum (EN, no subcategories). The selected sequelae were based on the study of Rosner et al.¹⁶

Incidence of illness

Estimates of incidence for yersiniosis were derived using notifications for the 2022 year (1,294) as a base. It was assumed that, since the commencement of laboratory-based notification in New Zealand,⁸ all isolates that tested positive for Yersinia spp. would result in notification. Factors for under-ascertainment of yersiniosis were taken as those for general AGI and were derived from a 2006 New Zealand study.¹⁵ Under-ascertainment factors were determined for: 1) the proportion of cases from whom a clinical sample was requested by a GP and not provided, 2) the proportion of cases of AGI for which no clinical sample was requested by a GP, and 3) the proportion of cases of AGI who did not present to a GP.

The proportion of cases hospitalised for yersiniosis was considered to be a subset of cases who visited a GP and was derived from the case-control study.

The incidence of sequelae (ReA and EN) was derived from a subset of the case-control study that was followed up at 3 months after the primary yersiniosis. The prevalence of self-reported ReA and EN in this cohort was used to estimate the incidence of these sequelae among all yersiniosis cases presenting to a GP. As with other studies of the burden of disease from yersiniosis, sequelae were considered to only occur in more serious cases of yersiniosis, those presenting to a GP.¹⁷

Deaths due to yersiniosis are extremely rare, with only a single fatality reported in the New Zealand notifiable disease database during the period 2003–2022 and none reported for the index year.¹³ The New Zealand case-fatality rate for the period 2003–2022 is 0.007% of notified cases. This case-fatality rate is lower than generally reported in the literature, with estimates in the range 0.03–1% of notified cases.^18–22 For the current study, it was assumed that the case fatality rate would be in the range 0.007–0.03% (uniform distribution), ranging from the New Zealand long-term mean to the bottom of the range reported internationally. As for sequelae, it was considered that fatalities would be associated with more serious cases of yersiniosis, those presenting to a GP and/or hospitalised. The very limited information on age at death for cases of yersiniosis internationally suggests they are older (50+ years).²³ This is consistent with the age of the single New Zealand fatality (55 years). Age at death for fatal cases of yersiniosis was modelled as a uniform distribution in the range 50–80 years.

Duration of illness

For primary yersiniosis, the duration of illness for cases presenting to a GP and hospitalised cases was derived from the case-control study. For a case, the duration of illness was taken as the maximum of the durations for the individual yersiniosis-related symptoms. The distribution of durations was fitted to a log-normal distribution, truncated at the minimum and maximum reported in the case-control study. The duration of illness for cases of yersiniosis not presenting to a GP was assumed to be the same as for general AGI in the New Zealand community.¹⁵

The case-control study did not follow cases with post-infectious sequelae to the resolution of symptoms. The duration of ReA was assumed to be the same as that resulting from other causes of AGI and was modelled as an exponential distribution with a mean of 0.608 years (222 days).^17,24,25

Little information is available on the duration of post-infectious EN. It has been reported that EN generally resolves spontaneously in 3–4 weeks.²⁶ A uniform distribution between 3 and 4 weeks was used to model the duration of EN.

Disability weights

Disability weights were based on those developed for the 2013 Global Burden of Disease Study²⁷ and are summarised in Table 1. It has been assumed that the disability associated with ReA will be similar to that due to osteoarthritis. No disability weight was found for EN, and it was assumed that this could be equated to “symptomatic other skin and subcutaneous diseases”.

View Table 1–3.

Burden of disease

The burden of disease due to yersiniosis in New Zealand in 2022 was estimated in terms of DALYs, as previously described.^6,28 For fatal cases, YLL was calculated from New Zealand tables of the residual life expectancy at the age of death.²⁹ In order to estimate the variability and uncertainty associated with DALY estimates, input variables were represented by statistical distributions (Table 2). The distributions were combined by simulation analysis using the Excel add-in @RISK (Lumivero, Denver, United States). Simulations were run for 100,000 iterations. Model outputs were presented as mean values and 95th percentile credible intervals.

Results

Illness incidence estimates and annual DALY burden of illness estimates for yersiniosis in New Zealand in 2022 are presented in Table 3. The uncertainty in the incidence and DALY estimates is shown as 95th percentile credible interval around the mean. The estimates in Table 3 are for the total amount of illness due to these diseases in New Zealand and will include illness due to transmission of the causative organisms by a variety of transmission routes.

Discussion

The burden of disease due to yersiniosis is predominantly due to the long duration of the gastrointestinal disease, with relatively minor contributions from sequelae. Preliminary data suggests that the burden of yersinosis in New Zealand is intermediate between the burdens of campylobacteriosis and listeria, and the burdens of salmonellosis and STEC (Shiga toxin E. coli) infection. While numbers of cases of yersiniosis reporting ReA and EN were quite low, the rates of these sequelae are consistent with rates found in other studies.

The study reported here represents the first assessment of the burden of yersiniosis in New Zealand, substantially based on New Zealand-specific data. In particular, information on the duration of the disease and the frequency of sequelae is New Zealand specific.

A range of methodological variations have been used in previous estimates of the burden of yersiniosis in New Zealand,^4,6,32,33 resulting in a range of burden estimates (54–111 DALYs). The current estimate of 119 DALYs is higher than previous New Zealand estimates, but not as proportionally higher as the substantial recent increases in the rate of notified yersiniosis in New Zealand would suggest.¹³ This relatively modest increase in the DALY estimate for yersiniosis is due to changes in the disability weights applied to the various disease states. For example, the first estimate of the burden of yersiniosis in New Zealand used a disability weight for moderate gastroenteritis of 0.39.⁶ Subsequent estimates used annualised disability weight, with a weight for moderate gastroenteritis of 0.015 for an incident of gastroenteritis of 10 days’ duration,^4,32,33 which equates to an incident-based disability weight of 0.55. The current study used disability weights revised for the 2013 Global Burden of Disease Study, with a weight for moderate gastroenteritis of 0.188.²⁷

While directly comparable estimates of the burden of disease due to other common enteric pathogens have not been published, work is underway, and initial findings suggest that the burden of disease due to yersiniosis is less than that due to campylobacteriosis and listeriosis, but greater than that due to salmonellosis or STEC infection.

Due to the long duration of moderate cases of yersiniosis (mean=18.5 days), these cases are a major determinant of the overall burden of disease. This extended duration has been noted in other studies. Ostroff et al.³⁴ reported a mean duration for a cohort of 67 cases of yersiniosis of 20 days. Rosner et al.¹⁶ reported a median duration of illness of 10 days (n=571). While not reporting an overall duration of disease, a study of 261 Dutch cases of yersiniosis noted that 34 had diarrhoea for 2–3 months and six experienced diarrhoea for up to 1 year.⁹

Compared to some other enteric diseases, sequelae to yersiniosis contribute a relatively small proportion (<10%) to the overall burden of disease in New Zealand. In comparison, some studies have estimated that 50% or more of the disease burden due to campylobacteriosis (ReA, Guillain-Barré syndrome and inflammatory bowel disease) and Shiga toxin-producing Escherichia coli infection (haemolytic uraemic syndrome, end-stage renal disease) is due to sequelae.^6,24 The more modest contribution of sequelae to the disease burden for yersiniosis is due to both their relative lack of severity and non-chronic nature, with EN generally lasting no more than a few weeks, while ReA may last for a few months.

ReA and EN have been consistently identified as sequelae occurring subsequent for enteric yersiniosis.^{9–11,16,34,35} In the current study, 138 cases of yersiniosis consented to re-interview after approximately 3 months, with seven cases reporting potential symptoms of ReA (5.1%, 95% confidence interval [CI] 2.1–10.2%) and three cases reporting potential symptoms of EN (2.2%, 95% CI 0.5–6.2). While the numbers of cases with self-reported sequelae were quite low, the incidence of these sequelae were generally in good agreement with other sources. A meta-analyses estimated that the proportion of yersiniosis cases that developed ReA was 3.4% (95% CI 0.8–13.7%).³⁵ Stolk-Engelaar and Hoogkamp-Korstanje⁹ reported 15 cases of arthritis out of 261 yersiniosis cases (5.7%, 95% CI 3.3–9.3%). Hannu et al.¹⁰ reported four cases of ReA from 33 cases of infection with Y. pseudotuberculosis O:3 (12.1%, 95% CI 3.4–28.2%). Rosner et al.¹⁶ reported symptoms of ReA in 12.4% of cases of yersiniosis and 4.8% of controls. The authors of this study concluded that this suggested a net rate of 6.6% of cases of yersiniosis with symptoms of ReA.

While less information is available on the incidence of EN subsequent to yersiniosis, Rosner et al.¹⁶ reported EN in 3.2% of cases of yersiniosis and just 0.1% of controls, and Stolk-Engelaar and Hoogkamp-Korstanje⁹ reported eight cases of EN out of 261 cases of yersiniosis (3.1%, 95% CI 1.3–6.0%).