Meta-analysis
This study systematically estimated the incidence of HZ, as well as the risk of PHN, recurrence, and hospitalisation, in China using the GRADE approach. Our analysis included approximately 25 million Chinese individuals from 12 studies, which is substantially more than that included in a previous review [5], which only included studies published in English (three from mainland China and seven from Taiwan) and those without any pooled HZ incidence.
In our study, the HZ incidence was 4.28‰, which is similar to that in multiple countries or regions of North America, Europe, and Asia-Pacific (3/1,000 to 5/1,000 person-years) [41]. We also demonstrated that the incidence in females was higher than that in males, that the incidence increased with age, and that the severity increased sharply in those older than 60 years, which are findings that are consistent with the results of other studies [42–46]. It was estimated that the annual number of HZ cases in China probably exceed 6 million according to the population of 1.41 billion at the end of 2021 [47], which is equivalent to the total of national legal infectious disease (6,233,537 cases) reported in 2021 [48]. Furthermore, as the population aging problem in China has entered a stage of rapid development, the proportion of individuals aged ≥ 65 years will increase from 6.8% in 2000 to 23.6% in 2050 [49, 50] and the number of HZ cases is expected to increase substantially.
It is generally believed that the occurrence of PHN is influenced by age and sex [51]. Due to the varying prevalence of disability and other underlying comorbidities, the risk of PHN varies from 5% to > 30% [41]. In our study, the risk of PHN in the patients with HZ was 12.6%. However, some studies gave an incorrect definition or no definition for PHN, as the best option for defining PHN would be clinically meaningful pain lasting for more than 90 days after rash onset [52]. We considered that PHN risk would be underestimated and that the actual situation is probably more serious. In addition to complications, the risks of hospitalisation and recurrence are indicators of HZ. The hospitalisation rate was 6.0/100,000 population in our study, which is similar to that in other countries (4–13.4/100,000 population) [53–55], and the risk of recurrence was 9.7%, which is slightly higher than that in some previous long-term follow-up studies in other countries (5–6%) [56, 57].
In summary, HZ has become a major public health problem with a significant health burden in China. This problem is more serious among the elderly, as the elderly population is growing rapidly. In addition to the severity and duration of pain associated with HZ, vaccination substantially reduces the incidence of HZ and incidence of PHN. RZV has been used in the European Union, United States, Japan, Canada, and Australia to prevent HZ or PHN in adults aged ≥ 50 years. Although RZV has been approved for use in China, it is not widely recommended, and both vaccination willingness and vaccination rate are low [58, 59]. Therefore, implementation of the immunisation strategy of the zoster vaccine for China should be considered to reduce the disease burden of HZ.
Evidence Quality Assessment
Accurately estimating incidence rates is difficult to achieve because HZ surveillance systems are lacking. Different study designs and case ascertainment may cause the differences observed in the incidence estimations. Some studies were community-based surveys; however, the response rates for these were not reported, which may lead to selection bias. Others who use big data from various sources may not include populations that are representative of the general population. Data collected from the Health Information System may exclude patients who did not seek medical care, and similar situations may occur in data from insurance databases or big-data platforms [41]. Furthermore, the retrospective study may lead to recall bias, which leads to both an overestimation and underestimation.
The evidence from observational studies is generally initially ‘low’ [25]. All of the studies included in this meta-analysis were observational, and even studies that are well-implemented generally receive a quality evidence rating of ‘low’. In addition to this possible bias, our assessment of the GRADE rating results of the quality of evidence of HZ incidence for all ages was ‘very low’, which indicates that the true incidence may be different from the estimated incidence. However, a ‘very low’ quality of evidence could still be strongly recommended [26] because our confidence has improved by findings of an incidence similar to that in other countries in our study. We believe that the burden of HZ in China is serious. Because the quality of evidence in the population aged 60 years or older is relatively improved, we have more confidence in that estimate.
Furthermore, in the process of assessing quality of evidence using the GRADE approach, we demonstrated that there was limited experience with GRADE in the evaluation of public health questions, especially for estimating incidences. The GRADE rating method needs to be clarified and more detailed regarding how to use it in the estimation of incidence. Second, for disease incidence studies, evidence is commonly derived from observational studies rather than randomized controlled trials. Even if the study is well-designed, using the GRADE approach immediately reduces the quality of evidence [60]. Further studies are needed to determine whether this is appropriate.