The prevalence and treatment pattern of clinically diagnosed pelvic organ prolapse: a Korean National Health Insurance Database-based cross-sectional study 2009–2015

The study aim was to evaluate the prevalence of pelvic organ prolapse using claim data of South Korea and to evaluate treatment patterns. The Korea National Health Insurance Corporation pay medical costs for most diseases. This study used Health Insurance Review & Assessment Service-National Inpatient Sample (HIRA-NIS) 2009–2015. Pelvic organ prolapse was defined by diagnostic code (N81.x). Of the approximately 4.5 million women included in HIRA-NIS 2009–2015, 10,305 women were selected as having pelvic organ prolapse, and the mean age of the pelvic organ prolapse group was 63.9 ± 0.2 years. The prevalence of pelvic organ prolapse was 180 ± 4 per 100,000 population in women older than 50 years old. In logistic regression analysis, constipation increased the prevalence of all pelvic organ prolapse (odds ratio, 4.04; 95% confidence interval, 3.52–4.63; P < 0.01). The number of women requiring pessary only and surgery only were 26 ± 2 per 100,000 population and 89 ± 1 per 100,000 population, respectively, for women over 50 years of age. The prevalence of pelvic organ prolapse was quite lower than in previous studies. Surgery peaked at approximately 70 years of age. Pessary increased dramatically among women after the age of 65.

100,000 population at all ages, respectively, and 26 ± 2 per 100,000 population, 89 ± 1 per 100,000 population, and 114 ± 2 per 100,000 population among women greater than 50 years of age, respectively. POP surgery was performed most often in women in their late 60 s and early 70 s, but the use of pessary was performed most often in women in their 70 s (Fig. 3). After 75 years of age, the use of pessary was higher than surgery. Of the total POP patients, 46% did not receive any special treatment, 44% underwent surgery, 9% used pessary, and 1% were treated with surgery and pessary.

Discussion
In this study, the prevalence of POP was 71 ± 1 per 100,000 population for all ages and 180 ± 4 per 100,000 population for women older than 50 years old. This rate is much lower than the prevalence reported in previous studies (2.9-41.1%) [5][6][7][8] . This finding is attributed to differences in POP definitions. Hendrix et al. diagnosed POP through direct physical examination, whereas other studies diagnosed POP using a survey about symptoms [5][6][7][8] . However, in our study, POP for the prevalence calculation was defined the case was diagnosed by the doctor in clinic. POP patients diagnosed by physical examination may include patients without discomfort. POP patients diagnosed based on symptoms may not exhibit POP by pelvic examination. In fact, the prevalence in this study (0.07%) was much lower than the prevalence of symptom-based studies (2.9-8.3%) [6][7][8] . This finding indicates that    the degree of discomfort to be treated is relatively low. The prevalence in this study is clinically more important than the prevalence reported by other studies because treating POP is determined by the degree of discomfort experienced by the patient 2 .
Our study reported that the prevalence of uterine prolapse was higher than that of cystocele. However, interpretation of this result requires caution. Our result represented exactly that the prevalence of diagnostic codes with uterine prolapse (Incomplete uterovaginal prolapse N81.2, Complete uterovaginal prolapse N81.3) was higher than that of diagnostic code with cystocele (Cystocele N81.1). However, cystocele or uterine prolapse is included in other diagnostic codes (Unspecified uterovaginal prolapse N81.4, Unspecified female genital prolapse N81.9). The ratios of cystocele or uterine prolapse in these codes are unknown. Therefore, this result should be interpreted with caution.
The rate of surgery in this study was 36 ± 0 per 100,000 population for all ages (89 ± 1 per 100,000 population for women greater than 50 years old), which was much lower than the rate of surgery (150 per 100,000 population) in the previous studies 11,12 . In addition, the rate of surgery in the previous studies (150 per 100,000 population) was greater compared with the prevalence of POP in this study (71 ± 1 per 100,000 population at all ages) 11,12 . This finding indicates that the prevalence of POP in previous studies is increased compared with the prevalence of our study. The exact reason is unknown, but we assumed the following information could explain this difference. First, the difference can result from a racial difference. Most of previous studies included Caucasians (81 ~ 97%) from the U.S.A., whereas the majority of our study population was Korean 11,12 . Therefore, white women may be at an increased risk for POP compared with Korean women. However, no significant difference was noted in the prevalence between Caucasians and Asians in previous studies 5,7 . Further studies adjusting environment, race, and country are needed. Second, economic differences between the two countries may explain the finding. The gross domestic product (GDP) for each person ($ 27,632) in South Korea is only 48% of the US GDP ($ 57,293) in 2015 13 . Due to the relatively low economic power of South Korea, patients with mild POP may have a tendency to receive less treatment. However, given that South Korea provides national health insurance, the surgery cost paid by the individual is relatively inexpensive. Therefore, it is unlikely the economic differences between the two countries cause the differences in POP prevalence. Third, there is a possible difference in the demographic composition between studies. The risks, such as parity, body weight, and constipation, may be different from those of previous studies 7 .
Pessary has few contraindication and is preferred for non-surgical treatment. Thus, 72% of US specialist clinicians choose pessary as the primary treatment for POP [14][15][16][17] . Pessary is successful in 53-76% of patients. However, the severity of POP, duration of use and type of pessary vary [18][19][20] . Despite the relatively high preference and success rate, few studies on how often pessary is used are available. In our study, pessary was used in 10% of all POP patients and 18.5% of all surgical or pessary treatments. In contrast, POP surgery was performed in 45% of all POP patients and peaked in patients approximately 70 years old. The use of pessary increases with age. In particular, the use of pessary in women 75 years or older was noted more often than any single surgery. (Fig. 3) The reason for this finding might be that the risk of surgery increases with age 21,22 . Considering that previous pelvic surgery is a risk for pessary failure and the reoperation rate of POP surgery was 29.2%, pessary should be used more often in patients with POP in their late 40 s to early 60 s 11,15 .
In our study, constipation was an important risk for POP. The results of previous studies are not consistent with this finding 5,7,23,24 . Hendrix et al. claimed that constipation is not a risk for POP, whereas most recent studies claimed that constipation is an important risk for POP 5,7,23,24 . Constipation might damage the pelvic floor (nerve and connective tissue) by increasing intra-abdominal pressure 7 . In support of this notion, one study reported that constipation in young adults caused POP 25 . However, in our study, the odds ratio of constipation in rectocele {OR 16.66 (13.76-20.17)} was significantly increased compared with the odds ratio of constipation in cystocele or uterine prolapse (1.85-2.12). Although constipation worsens POP, rectocele potentially caused constipation unlike cystocele or uterine prolapse. Bozkurt et al. reported that rectocele is a risk of constipation 26 . Both diseases (constipation and rectocele) are likely to exhibit negative synergy with each other. Further studies of the causal relationship between rectocele and constipation are needed.
Our study has some limits. First, our study could not distinguish vault prolapse from uterine prolapse. Given that the HIRA-National Inpatient Sample (HIRA-NIS) used in our study contains one-year sample data, we could not confirm the presence of hysterectomy prior to 1 year. Second, our data did not confirm the stage for cystocele and rectocele. Therefore, the prevalence according to stage was not confirmed. Third, our study included no data on parity or occupation. Therefore, our study could not adjust these factors. However, given that the primary purpose of our study was to determine the prevalence of POP, we did not experience problems in obtaining the prevalence of POP.
In conclusion, the prevalence of POP was 180 ± 4 per 100,000 population among women over 50 years old, which was quite lower than that noted in previous studies. Surgery peaked approximately 70 years old. The use of pessary has increased dramatically in women older than 65 years, and this procedure is the most commonly used treatment for women over 75 years old.

Materials and Methods
Study Settings and Participants. The Republic of Korea provides medical insurance service {(the National Health Insurance Corporation (NHIC)} to almost all Koreans living in the Republic of Korea 27 . Given that the NHIC offers medical insurance services for most diseases except for special cases, such as cosmetic surgery, it contains copious medical information, such as gender, age, low-income households group, diagnosis name, surgery name, and prescription history 27 . The Health Insurance Review & Assessment Service (HIRA) is an organization that evaluates the medical expenses charged by medical institutions in a neutral manner. The HIRA decides whether the costs are appropriate and suggests that the NHIC pay these costs. Therefore, the HIRA shares a significant portion of NHIC data 27 .
The HIRA-NIS is annual sample data using a stratifed randomized sampling method provided by the HIRA for medical research. The HIRA-NIS perform sampling in each year. Therefore, the sample members of each year are not the same. Per the extraction method, the HIRA-NIS extracts data from 13% of patients who were admitted during a one-year period and 1% of patients who were not admitted during a one-year period. This study used HIRA  27 .
The surgical and treatment codes used were the medical care benefits of the health insurance 2016 edition, and the diagnostic code used was the Korean Standard Classification of Diseases, 7th Edition (KCD-7), which was modified from the International Statistical Classification of Diseases and Related Health Problems, 10 th edition (ICD-10).
As To determine the disease risk, women with chronic obstructive pulmonary disease (J44.x) and constipation (K59.0) were defined as having the disease when they had more than applicable diagnosis code. We have defined women who received the livelihood program as the low SES group. Statistical Analysis. All statistical analyses in this study were performed using statistical program R version 3.3.2 (The R Foundation for Statistical Computing, Vienna, Austria). All statistical calculations were performed using two-tailed tests and were assumed to be statistically significant if the p-value was less than 0.05. The weighted t-test was utilized for the mean comparison of continuous variables, and the chi-square test was utilized for the comparison of categorical variables. The weighted logistic regression method was utilized to calculate the risk of multiple variables.
Ethics. Given that this study uses data anonymized by a third party, it is not subject to the Institutional Review Board (IRB) under the South Korea's Bioethics and Safety Act.