Introduction

The prevalence of type 2 diabetes mellitus is rapidly increasing, with the World Health Organization predicting that the number of cases worldwide will double by the year 2030 (http://www.who.int/diabetes/publications/en/screening_mnc03.pdf). Diabetes is often associated with cardiovascular problems such as ischemic heart disease [1], cerebrovascular accidence [2, 3], and peripheral vascular diseases [4]. It is also associated with an increased risk of suffering from renal complications [5] and premature mortality [4]. This increased risk of mortality in type 2 diabetes patients is nearly twice that of healthy controls [6], thus making the early identification of both type 2 diabetes and those predisposed toward contracting type 2 diabetes an important issue.

Perianal abscess is a benign anorectal disease and a common surgical condition with most perianal abscess patients being diagnosed during their fourth and fifth decades of life [7–9]. Its prevalence is difficult to estimate as the majority of patients with anorectal symptoms do not seek medical attention. While it has been characterized as the acute phase of a perianal suppuration process originating from an infected anal crypt gland [10], other factors producing chronic systemic inflammation, such as those arising on account of smoking, have been reported to be risk factors for perianal abscess [11], thus suggesting that this locally defined condition has systemic corollaries. Perianal abscess may also be the initial presentation of immunocompromised patients, such as those infected with human immunodeficiency virus [12] or suffering from an occult systemic disease [13].

Although type 2 diabetes is regarded as an inflammatory disease [14], previous studies have found that abnormal energy production might cause dysfunctions in the lymphomononuclear cells and immune systems of diabetic and prediabetic patients [15]. This may help explain the association between perianal abscess and its common complication, Fournier’s gangrene, which is a potentially life-threatening necrotizing infection with diabetes mellitus as its most common predisposing factor [16]. However, it is unclear whether perianal abscess indicates an increased risk of subsequent development of diabetes, or whether no such relationship exists. Therefore, using a population-based dataset, this study aimed to explore the risk of developing type 2 diabetes subsequent to diagnosis with perianal abscess.

Methods

Database

We used data sourced from the Longitudinal Health Insurance Database (LHID2000). Taiwan launched its National Health Insurance (NHI) program in 1995 to provide affordable health care for all the residents of Taiwan. The LHID2000 was created by the Taiwan National Health Research Institutes and includes all the registration files and medical claims for 1,000,000 NHI enrollees randomly selected from the 2000 Registry of NHI Beneficiaries (n = 23.72 million). The Taiwan National Health Research Institutes has validated the representativeness of the gender distribution of the LHID2000 relative to the whole population of NHI enrollees. Numerous studies utilizing data sourced from the LHID2000 have been published in internationally peer-reviewed journals [17, 18].

This study was exempt from full review by an institutional review board as the dataset utilized in the investigation only consisted of de-identified secondary data released without restrictions for research purposes.

Study sample

This study included a study group and a comparison group. When constructing our study group, we first identified 1,754 patients who were hospitalized or visited ambulatory care centers with a diagnosis of perianal abscess (ICD-9-CM code 566, Abscess of anal and rectal regions) between January 1, 2002, and December 31, 2004. We considered their first hospitalization or ambulatory care visit for the treatment of perianal abscess as their index date. We then excluded patients aged less than 18 years (n = 36). We also excluded patients who were either diagnosed with diabetes (ICD-9-CM code 250) prior to their index date or within 7 days following their index date (n = 299). We excluded patients receiving a diagnosis of diabetes within the first 7 days following their index date to mitigate the chance of introducing the effects of a surveillance bias into our study. It is possible that some patients were unaware of their diabetes status and only became aware of their disease on account of having received a physical examination associated with the treatment of perianal abscess. Thus, their inclusion would have biased our results. Ultimately, 1,419 adult patients with perianal abscess were included in the study group.

As for the comparison group, we selected the remaining NHI beneficiaries registered in the LHID2000 who did not receive a diagnosis of perianal abscess since the initiation of the NHI in 1995. We further excluded subjects aged less than 18 years. We then randomly selected 7,095 subjects (five comparison patients for every patient with perianal abscess) matched on sex, age (<40, 40–49, 50–59, and ≥60 years), and year of the study patient’s index date using the SAS proc surveyselect program (SAS System for Windows, version 8.2, SAS Institute Inc., Cary, NC). For comparison patients, we assigned their first use of health care in the index year as their index date, regardless of whether it occurred in an inpatient or ambulatory setting. Similarly, both subjects diagnosed with diabetes prior to their index date and within 7 days following their index date were excluded.

We individually tracked each subject (n = 8,514) for a 5-year period beginning with their index date to identify those patients who had received a diagnosis of diabetes. Since the National Health Insurance Research Database allows us to trace all use of medical services for all enrollees, all sampled patients could be followed throughout the study period. During the 5-year follow-up period, 1,052 (12.4 %) sampled patients died; 208 (14.7 %) were from the study group and 844 (11.9 %) were from the comparison group.

Statistical analysis

We used Pearson χ 2 tests to compare the geographic location (Northern, Central, Eastern, and Southern Taiwan) and urbanization level (five strata, with level 1 referring to the “most urbanized” and level 5 referring to the “least urbanized” communities) of the patient’s residence, monthly income, and selected medical comorbidities measured at baseline (hypertension, coronary heart disease (CHD), hyperlipidemia, renal disease, obesity, and alcohol abuse/alcohol dependence syndrome) between perianal abscess and comparison patients. We only included a comorbidity if the diagnosis appeared in an inpatient claim or was coded for at least twice in ambulatory care claims 1 year before the index date.

The Kaplan–Meier method and log-rank test were performed to generate survival curves and compare the 5-year diabetes-free survival rate between perianal abscess patients and the comparison group. Furthermore, stratified Cox proportional hazards regressions (stratified by age, sex, and the year of index date) were carried out in order to evaluate the association between perianal abscess and being subsequently diagnosed with diabetes within 5 years after censoring cases that died from nondiabetes causes and adjusting for selected medical comorbidities. The hazard ratios (HR) and 95 % confidence intervals (95 % CI) were used to estimate the risk of diabetes. A two-sided p value of <0.05 was considered statistically significant in this study.

Results

Table 1 presents the distribution of demographic characteristics and comorbid medical disorders measured at baseline among perianal abscess and comparison patients. Of the 8,514 sampled patients, the overwhelming majority (82.9 %) were males with 43 % aged less than 40 years. There were no significant differences found in the distribution of the geographic location (p = 0.355), urbanization level (p = 0.347) of the patient’s residence, or monthly income (p = 0.446) between patients with perianal abscess and sex- and age-matched comparison patients. However, patients with perianal abscess were more likely than the comparison group to suffer from the following comorbidities: hypertension (p < 0.001), hyperlipidemia (p < 0.001), CHD (p < 0.001), and obesity (p = 0.022).

Table 1 Demographic characteristics for the sampled patients in Taiwan, stratified by the presence/absence of perianal abscess, 2002–2004 (n = 8,514)
Full size table

During the 5-year follow-up period, patients with perianal abscess had a higher incidence of diabetes than comparison patients (13.9 vs. 7.8 %). Figure 1 presents the diabetes-free survival curves generated by the Kaplan–Meier method and shows that patients with perianal abscess had significantly lower 5-year diabetes-free survival than comparison patients (log-rank test: 17.33; p < 0.001).

Fig. 1
figure 1

Diabetes-free survival rates for patients with perianal abscess and the comparison group

Full size image

Table 2 shows the incidence rate of diabetes per 100 person-years. Of the total 8,514 sampled subjects, the incidence rate of diabetes per 100 person-years was 1.87 (95 % CI = 1.74–2.01); the rate among patients with perianal abscess was 3.00 (95 % CI = 2.60–3.43) and 1.65 (95 % CI = 1.52–1.79) among comparison patients. Table 2 also presents the crude HR for diabetes. Stratified Cox proportional hazards analysis (stratified by age, sex, and the year of index date) revealed that relative to comparison patients, patients with perianal abscess had a crude HR of 1.92 (95 % CI = 1.61–2.28, p < 0.001).

Table 2 Crude hazard ratios for diabetes among the sampled patients during the 5-year follow-up period starting from the index date
Full size table

Table 3 shows the covariate-adjusted diabetes HRs. After censoring cases that died from nondiabetes causes and adjusting for patient’s geographic location, urbanization level, monthly income, hypertension, CHD, hyperlipidemia, obesity, and alcohol abuse/alcohol dependence syndrome at baseline, stratified Cox proportional hazards analysis (stratified by age, sex, and the year of index date) revealed that patients with perianal abscess were more likely to have been diagnosed with diabetes relative to comparison patients (HR = 1.80, 95 % CI = 1.50–2.16, p < 0.001) during the 5-year follow-up period.

Table 3 Covariate-adjusted hazard ratios for diabetes among the sampled patients during the 5-year follow-up period (n = 8,514)
Full size table

Discussion

In this pioneering study, we found that patients were more likely to develop type 2 diabetes mellitus within 5 years after being diagnosed with perianal abscess than patients that did not receive a diagnosis of perianal abscess. Type 2 diabetes mellitus is characterized by hyperglycemia as a result of insulin resistance. Although the pathogenesis is still poorly understood, both genetic and environmental factors have been shown to contribute toward its development. Type 2 diabetes mellitus is regarded as a serious disease due to its metabolic complications, number of attributed deaths, and economic burden [19]. Some of the risk factors for type 2 diabetes mellitus are modifiable, such as maintaining a sedentary lifestyle, being overweight or obese, and being diagnosed with prediabetes, dyslipidemia, or hypertension [20]. The importance of early identification becomes further apparent as it has been demonstrated that the incidence of type 2 diabetes mellitus can be reduced among high-risk individuals by avoiding a Western diet, increasing activity levels, and losing weight [21].

Although the cryptoglandular obstruction theory for perianal abscess formation has been widely accepted [22, 23], neither have the details of the mechanism been fully elucidated, nor can the model be used to adequately explain the observed epidemiological phenomenon surrounding perianal abscess to date. For example, it can still not be explained why more males suffer from perianal abscess formation than women [7–9, 24] or why perianal abscess is frequently observed in the patients with immune-compromised conditions such as leukemia [25], post-bone marrow transplantation [26], and autoimmune neutropenia [27]. Finally, it is still unknown why the chronic systemic inflammation arising from smoking can also be regarded as a risk factor for perianal abscess [11].

In agreement with previous work, this study found perianal abscess to be male gender predominant and to occur more frequently among individuals suffering from metabolic disorders such as obesity, hypertension, and hyperlipidemia. Furthermore, coronary heart disease was noted to be more prevalent in patients with perianal abscess than comparison patients. This observation contributes to the understanding of perianal abscess as a systemic disease.

We further investigated the relationship between perianal abscess and a subsequent diagnosis with type 2 diabetes mellitus. During the 5-year follow-up period, patients with perianal abscess had a higher incidence of diabetes than comparison patients (13.9 vs. 7.8 %). The incidence rate of diabetes per 100 person-years was 3.00 (95 % CI = 2.60–3.43) for patients with perianal abscess and 1.65 (95 % CI = 1.52–1.79) for those without. Patients with perianal abscess had significantly lower diabetes-free survival than the comparison group.

Obesity, hypertension, and hyperlipidemia are well-known risk factors for type 2 diabetes mellitus. Because the incidence of these three comorbidities was also higher among patients with perianal abscess, we further controlled for these factors to explore the relationship between the perianal abscess and type 2 diabetes mellitus. After adjusting for obesity, hypertension, hyperlipidemia, and coronary heart disease at baseline, stratified Cox proportional hazards analysis (stratified by age, sex, and the year of index date) revealed that patients with perianal abscess had a higher rate of type 2 diabetes mellitus (HR = 1.80, 95 % CI = 1.50–2.16, p < 0.001).

These results contribute to the classification of perianal abscess as an independent risk factor for development of type 2 diabetes mellitus. One possible mechanism underpinning this association might involve the contribution of low-grade inflammation, which could increase the risk of both metabolic syndrome and diabetes [28].

The strength of our study lies in its longitudinal database and large population size. Nevertheless, the findings of this study need to be interpreted with awareness of one major limitation. The dataset used in this study lacks information on body mass index, diet, physical activity level, smoking, positive family history of diabetes, and alcohol consumption. These factors have been demonstrated to be associated with an increased risk of type 2 diabetes, and therefore, their exclusion from our study may have biased our results.

In conclusion, this study found that patients with perianal abscess had a higher chance of developing type 2 diabetes mellitus during a 5-year follow-up period. The results of this study highlight a need for clinicians dealing with perianal abscess patients to be alert to the possible development of diabetes and to counsel their patients on making lifestyle changes to reduce the impact of modifiable risk factors.