Prevalence and 5-year trend of incidence for medical illnesses after the diagnosis of bipolar disorder: A nationwide cohort study

Abstract

Objective:

Medical comorbidities are prevalent in patients with bipolar disorder. Evaluating longitudinal trends of the incidence of medical illnesses enables implementation of early prevention strategies to reduce the high mortality rate in this at-risk population. However, the incidence risks of medical illnesses in the early stages of bipolar disorder remain unclear. This study investigated the incidence and 5-year trend of medical illnesses following bipolar disorder diagnosis.

Methods:

We identified 11,884 patients aged 13–40 years who were newly diagnosed as having bipolar disorder during 1996–2012 and 47,536 age- and sex-matched controls (1:4 ratio) who represented the general population from Taiwan’s National Health Insurance Research Database. We estimated the prevalence and incidence of individual medical illnesses yearly across the first 5 years after the index date. The adjusted incidence rate ratio was calculated to compare the occurrence of specific medical illnesses each year between the bipolar disorder group and control group using the Poisson regression model.

Results:

Apart from the prevalence, the adjusted incidence rate ratios of most medical illnesses were >1.00 across the first 5-year period after bipolar disorder diagnosis. Cerebrovascular diseases, ischaemic heart disease, congestive heart failure, other forms of heart disease, renal disease and human immunodeficiency virus infection exhibited the highest adjusted incidence rate ratios during the first year. Except for that of renal disease, the 5-year trends of the adjusted incidence rate ratios decreased for cerebrovascular diseases, cardiovascular diseases (e.g. ischaemic heart disease, other forms of heart disease, and vein and lymphatic disease), gastrointestinal diseases (e.g. chronic hepatic disease and ulcer disease) and communicable diseases (e.g. human immunodeficiency virus infection, upper respiratory tract infection and pneumonia).

Conclusion:

Incidence risks of medical illnesses are increased in the first year after bipolar disorder diagnosis. Clinicians must carefully evaluate medical illnesses during this period because the mortality rates from medical illnesses are particularly high in people with bipolar disorder.

Keywords

Bipolar disorder medical comorbidity incidence incidence rate ratio trend

Introduction

Bipolar disorder is a severe mental illness that usually manifests as episodic mood swings, cognitive impairment and socio-occupational dysfunction (Baune and Malhi, 2015; Grande et al., 2016; Jauhar et al., 2019). Compared with the general population, individuals with bipolar disorder are at 1.5–3 times higher risk of developing various medical illnesses (Correll et al., 2017; Crump et al., 2013; Firth et al., 2019; Forty et al., 2014; Goldstein et al., 2015; Leboyer et al., 2012). Furthermore, among individuals with bipolar disorder, medical comorbidities markedly increase the severity of mood symptoms, rate of suicide attempts, chance of relapse and rehospitalisations, and number of life-years lost (Crump et al., 2013; Forty et al., 2014; Kemp et al., 2010, 2014; Leboyer et al., 2012; Plana-Ripoll et al., 2019). Numerous studies have indicated the bidirectional association between mental and physical dimensions during the progression of bipolar disorder (Forty et al., 2014; Kemp et al., 2010; Kemp et al., 2014; Leboyer et al., 2012) and have proposed the incorporation of medical illnesses into staging models of bipolar disorder to determine their chronicity and complexity in terms of bipolar illness (Berk et al., 2017; Leboyer et al., 2012; Morris et al., 2019; Shah et al., 2020). Implementing this proposal necessitates an accurate description of the onset of medical illnesses and their longitudinal course during the progression of bipolar disorder. In particular, research focusing on the occurrence of medical illnesses during the early stages of bipolar disorder would be helpful in guiding early intervention strategies against premature deaths in this at-risk group. Several studies have investigated the prevalence of medical illnesses among young patients with early-onset bipolar disorder (Evans-Lacko et al., 2009; Goldstein et al., 2015; Jerrell et al., 2010) and have revealed that medical comorbidities are already prevalent even in patients with bipolar disorder during adolescence and young adulthood. Incidence is a measure of the frequency at which new cases occur over a specified period; accordingly, evaluating longitudinal trends for the incidence of medical illnesses can provide insights into the progression of such illnesses. However, data regarding the incidence of medical illnesses during the early stages of bipolar disorder are lacking.

In regard to the knowledge gap and clinical merits, we conducted this study to investigate the incidence of various medical illnesses in a large bipolar disorder cohort using a nationwide population-based claims database. We randomly selected age- and sex-matched controls (at a 1:4 ratio) from a representative cohort of the general population in Taiwan. The trends of the incidence rate ratios (IRRs) of medical illnesses were evaluated in each of the first 5 years following bipolar disorder diagnosis to determine the incidence risks of such illnesses during the early stages of this disorder.

Methods

The Taipei City Hospital Research Ethics Committee approved the present study protocol (TCHIRB-10708113-E) and waived the need for informed consent due to the deidentified and retrospective nature of data. Data that could be used to identify beneficiaries and medical care providers were encrypted by the Taiwan National Health Insurance (NHI) Administration.

Data sources

Data used in this study were derived from the Psychiatric Inpatient Medical Claims (PIMC) database – a subset of the Taiwan National Health Insurance Research Database (NHIRD) – for the period between 1 January 1996 and 31 December 2012. In Taiwan, the NHI programme was implemented in 1995 and provides comprehensive medical care for nearly 98% of the 23,000,000 people in Taiwan. The NHIRD contains the registration files and medical claims data of beneficiaries covered by the NHI programme. Information available in the NHIRD includes detailed diagnostic codes according to the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM).

The PIMC database includes data on psychiatric inpatients with ICD-9-CM codes of 290–319 for the period between 1 January 1996 and 31 December 2012. Each year, the NHI Administration conducts randomised reviews of medical records to verify the diagnoses. In addition, the Taiwan Joint Commission on Hospital Accreditation oversees the accreditation of hospitals contracted with the NHI. The accreditation for qualified psychiatric services requires board-certified psychiatrists to diagnose patients. Therefore, the PIMC database has acceptable diagnostic coding accuracy for epidemiological research. To date, the PIMC database has been the data source for numerous epidemiological studies (Chen et al., 2019b, 2020b).

Identification of newly diagnosed bipolar disorder cohort

To identify the bipolar disorder cohort, we first selected a cohort of patients from the PIMC database who had at least one psychiatric admission for bipolar disorder (ICD-9-CM codes: 296.0–296.16, 296.4–296.81 and 296.89) between 1 January 1998 and 31 December 2012 (e-Figure 1 in the Supplemental Material). Subsequently, we excluded patients who had psychiatric admission between 1 January 1996 and 31 December 1997 to ensure that we included only patients who had their first incident of bipolar disorder that required psychiatric hospitalisation during the study period. We then excluded patients who ever received a diagnosis of schizophrenia (ICD-9-CM codes: 295.**) between 1 January 1996 and 31 December 2012. By adopting the definition of early-onset bipolar disorder (i.e. onset before the age of 40 years) proposed recently by the International Society for Bipolar Disorders Task Force (Sajatovic et al., 2015), we restricted the patient age to 13–40 years and eventually identified 11,884 patients who received a diagnosis of early-onset bipolar disorder between 1 January 1998 and 31 December 2011 (i.e. the study cohort).

To minimise the confounding effects of age and sex, for each patient with bipolar disorder, we randomly selected four age-matched (±0 year) and sex-matched controls from the NHIRD that represented the general population in Taiwan (e-Figure 1 in the Supplemental Material). The controls were assigned the same index date (i.e. the date of the first diagnosis of bipolar disorder) as their corresponding case. To confirm that the controls were alive before the corresponding index dates, each control was required to have at least one claims record after the corresponding index date. Accordingly, a total of 11,884 patients with bipolar disorder were matched with 47,536 controls.

Demographic and clinical variables

We obtained demographic and clinical variables, namely, sex, age, urbanisation level of residence location, employment and Charlson Comorbidity Index (CCI) score, at the index date for both patients and age- and sex-matched controls. Urbanisation was categorised into the following five levels, which are used specifically in Taiwan: Level 1 (highly urbanised area), Level 2 (moderately urbanised area), Level 3 (township area), Level 4 (subrural area) and Level 5 (rural area). The CCI was used to categorise the severity of medical comorbidities (assigned with scores of 0, 1 or ⩾2) according to the ICD-9-CM codes for primary and other diagnoses in the NHIRD (Charlson et al., 1987; Romano et al., 1993). The CCI comprised 17 medical illnesses within 1 year prior to the index date, namely, myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, chronic pulmonary disease, connective tissue disease, ulcer disease, mild liver disease, diabetes, diabetes with end organ damage, hemiplegia, moderate or severe renal disease, any tumour, moderate or severe liver disease, metastatic solid tumour and acquired immune deficiency syndrome.

Variables of medical illnesses

In this study, we categorised medical illnesses into the following seven major organ systems: cardiovascular, cerebrovascular, respiratory, gastrointestinal, renal, endocrine/metabolic and musculoskeletal/integument. We also investigated the prevalence and incidence of human immunodeficiency virus (HIV) infection and cancer because these illnesses are increasing in young patients with bipolar disorder (Chen et al., 2019a; Hung et al., 2014). Detailed information on ICD-9-CM codes for the medical illnesses considered in this study is listed in e-Table 1 in the Supplemental Material.

Statistical analysis

We first computed the prevalence of each medical illness within the first year after the diagnosis of bipolar disorder. The risks of prevalent medical illnesses between the bipolar disorder cohort and control cohort were initially compared using a univariable conditional logistic regression analysis (Table 2). Subsequently, we conducted a multiple conditional logistic regression analysis to estimate the risk of each prevalent medical illness within 1 year after bipolar disorder. The backward stepwise strategy was used to select variables with p < 0.001 in the final model (Table 3).

Next, we calculated the first-year incidence of each medical illness after the diagnosis of bipolar disorder (Table 4). Consider, for example, hypertension: To estimate the incidence of hypertension, we first excluded patients who had hypertension prior to bipolar disorder diagnosis (i.e. between the date of the first diagnosis of bipolar disorder and 1 January 1996). Subsequently, we computed the incidence of hypertension after the diagnosis of bipolar disorder as the number of new cases with hypertension divided by the contributed person-years of the cohort without preexisting hypertension. We used the same method to calculate the incidence of each medical illness after the index date in the matched control cohort. The IRR was calculated as the ratio of the incidence of a specified medical illness between the bipolar disorder cohort and control cohort to determine whether the occurrence of this medical illness in the bipolar disorder cohort differed from that in the control cohort. To reveal the trends of the incidence and IRRs across the 5-year period after the diagnosis of bipolar disorder, we calculated the IRRs of each medical illness in the first, second, third, fourth and fifth year after the date of the first diagnosis of bipolar disorder (Figure 1). We applied the Poisson regression model to estimate the adjusted IRRs (aIRRs) and their 95% confidence intervals (Szklo and Nieto, 2019). A person–time value was registered for each patient and control from the index date until the occurrence of a specified disease incidence or mortality. Besides, we respectively conducted the Poisson regression model to evaluate the aIRRs of each medical illness between the first year and subsequent second, third, fourth and fifth year after the diagnosis of bipolar disorder. A p value of <0.001 was considered statistically significant in this study.

Figure 1.

Adjusted incidence rate ratio (95% CI) of various medical illnesses within the first 5 years after the index date among patients with newly diagnosed bipolar disorder compared with age- and sex-matched controls.

Sensitivity analysis

Because patients with new-onset bipolar disorder could be younger than 30 years, and approximately 38% of the patients included in this study were aged above 30 years when they received their first diagnosis of bipolar disorder, we performed a sensitivity analysis using a subgroup of patients aged 13–29 years to estimate the incidence and aIRRs of each medical illness within 1 year after the index date in the patient cohort and control cohort. The study flow diagram is illustrated in e-Figure 2 in the Supplemental Material. Furthermore, to consolidate our results for patients with a first diagnosis of bipolar disorder, we performed another sensitivity analysis using a subgroup of patients who were entered into the cohort between 1 January 2001 and 31 December 2011 to allow for a 5-year washout period (i.e. between 1 January 1996 and 31 December 2000). The study flow diagram is presented in e-Figure 3 in the Supplemental Material. All statistical analyses were conducted using SAS v9.4 (SAS Institute, Cary, NC, USA).

Results

Demographic and clinical characteristics

Table 1 presents the demographic and clinical characteristics of the patient cohort and control cohort. The male-to-female sex ratio in the study cohort was 1.30. Approximately 60% of the patients in the study cohort were newly diagnosed as having bipolar disorder before the age of 30 years. Compared with the control group, the patients tended to live in highly urbanised area, be unemployed and have significantly higher CCI scores.

Table 1.

Demographic and clinical characteristics of patients with newly diagnosed bipolar disorder and age- and sex-matched controls at the index (corresponding) date.

	Patients(N = 11,884)	Controls(N = 47,536)	Unadjusted odds ratio^a	95% CI	P value
	n (%)	n (%)
Sex
Male	6712 (56.5)	26,848 (56.5)	Reference	–	–
Female	5172 (43.5)	20,688 (43.5)	–	–	–
Age, mean±SD (years)
13–19	1532 (12.9)	6128 (12.9)	Reference
20–29	5787 (48.7)	23,148 (48.7)	–	–	–
30–40	4565 (38.4)	18,260 (38.4)	–	–	–
Urbanisation^b
Level 1	7126 (56.0)	15,920 (33.5)	Reference	–	–
Level 2	1782 (15.0)	16,532 (34.8)	0.32	[0.30, 0.33]	<0.001
Level 3	732 (6.2)	6478 (13.6)	0.33	[0.30, 0.35]	<0.001
Level 4	887 (7.5)	5330 (11.2)	0.46	[0.43, 0.50]	<0.001
Level 5	1357 (11.4)	3110 (6.5)	0.98	[0.93, 1.04]	0.550
Employment
Yes	5127 (43.1)	25,464 (53.6)	Reference	–	–
No	6757 (56.9)	22,072 (46.4)	1.40	[1.35, 1.45]	<0.001
Charlson Comorbidity Index^b
0	9243 (77.8)	43,780 (92.1)	Reference	–	–
1	1927 (16.2)	3133 (6.6)	2.19	[2.08, 2.30]	<0.001
⩾2	714 (6.0)	623 (1.3)	3.06	[2.84, 3.31]	<0.001

CI: confidence interval.

Estimated using univariate conditional logistic regression.

At the index (corresponding) date. The level of urbanisation was categorised as Level 1 (highly urbanised area), Level 2 (moderately urbanised area), Level 3 (township area), Level 4 (subrural area) and Level 5 (rural area).

Medical illnesses within 1 year after bipolar disorder diagnosis

Table 2 shows the prevalence of medical illnesses within the first year after bipolar disorder diagnosis. Compared with the control cohort, the patients exhibited an increased risk of numerous medical illnesses across all major organ systems. Furthermore, the multiple conditional logistic regression analysis results demonstrated that the patient cohort had a significantly higher risk of medical illnesses among six organ systems in the first year after diagnosis (Table 3). The risk of HIV infection also increased for patients within 1 year following bipolar disorder diagnosis.

Table 2.

Prevalence of medical illnesses within 1 year after the index date among patients with newly diagnosed bipolar disorder and age- and sex-matched controls.

	Patients(N = 11,884)	Controls(N = 47,536)	Unadjusted odds ratio^a	95% CI	P value
	n (%)	n (%)
Cardiovascular diseases
Hypertension	379 (3.2)	633 (1.3)	1.90	[1.72, 2.11]	<0.001
Ischaemic heart disease	153 (1.3)	171 (0.4)	2.38	[2.03, 2.79]	<0.001
Congestive heart failure	46 (0.4)	37 (0.1)	2.74	[2.06, 3.63]	<0.001
Other forms of heart disease	520 (4.4)	569 (1.2)	2.45	[2.25, 2.68]	<0.001
Diseases of arteries, arterioles and capillaries	74 (0.6)	117 (0.25)	1.94	[1.55, 2.44]	<0.001
Diseases of veins and lymphatics	547 (4.6)	835 (1.8)	2.03	[1.86, 2.21]	<0.001
Cerebrovascular diseases	206 (1.7)	104 (0.2)	3.36	[2.93, 3.86]	<0.001
Respiratory diseases
URI	2162 (18.2)	8108 (17.1)	1.06	[1.02, 1.12]	0.009
Pneumonia	322 (2.7)	473 (1.0)	2.06	[1.84, 2.30]	<0.001
Asthma	368 (3.1)	671 (1.4)	1.80	[1.62, 1.99]	<0.001
Chronic bronchitis	177 (1.5)	381 (0.8)	1.60	[1.38, 1.85]	<0.001
COPD	425 (3.6)	1051 (2.2)	1.46	[1.32, 1.60]	<0.001
Tuberculosis	47 (0.4)	56 (0.1)	2.30	[1.73, 3.06]	<0.001
Gastrointestinal diseases
Chronic hepatic disease	1055 (8.9)	1331 (2.8)	2.33	[2.19, 2.48]	<0.001
Ulcer	1313 (11.1)	1875 (3.9)	2.19	[2.07, 2.32]	<0.001
Irritable bowel syndrome	327 (2.8)	516 (1.1)	1.97	[1.76, 2.20]	<0.001
Renal diseases
Chronic kidney disease and renal failure	52 (0.4)	78 (0.2)	2.01	[1.53, 2.63]	<0.001
Endocrine/metabolic diseases
Diabetes mellitus	274 (2.3)	373 (0.8)	2.15	[1.91, 2.42]	<0.001
Hyperlipidaemia	404 (3.4)	685 (1.4)	1.89	[1.71, 2.08]	<0.001
Musculoskeletal/integumentary diseases
Connective tissue disease	166 (1.4)	259 (0.5)	1.97	[1.69, 2.29]	<0.001
Atopic dermatitis	176 (1.5)	532 (1.1)	1.25	[1.08, 1.45]	0.004
HIV infection	31 (0.3)	28 (0.1)	2.63	[1.85, 3.74]	<0.001
Cancer	87 (0.7)	186 (0.4)	1.60	[1.29, 1.97]	<0.001

CI: confidence interval; COPD: chronic obstructive pulmonary disease; HIV: human immunodeficiency virus; URI: upper respiratory tract infection.

Estimated using univariate conditional logistic regression.

Table 3.

Multiple conditional logistic regression of prevalent medical illnesses within 1 year after the index date among patients with newly diagnosed bipolar disorder and age- and sex-matched controls.

	Adjusted odds ratio^a	95% CI	P value
Cardiovascular diseases
Hypertension	1.26	[1.13, 1.40]	<0.001
Other forms of heart disease	1.77	[1.62, 1.94]	<0.001
Diseases of veins and lymphatics	1.55	[1.42, 1.70]	<0.001
Cerebrovascular diseases	2.06	[1.79, 2.37]	<0.001
Respiratory diseases
Pneumonia	1.51	[1.34, 1.69]	<0.001
Asthma	1.40	[1.26, 1.56]	<0.001
Tuberculosis	1.64	[1.23, 2.18]	<0.001
Gastrointestinal diseases
Chronic hepatic disease	1.69	[1.58, 1.81]	<0.001
Ulcer	1.63	[1.54, 1.74]	<0.001
Irritable bowel syndrome	1.34	[1.20, 1.50]	<0.001
Endocrine/metabolic diseases
Diabetes mellitus	1.41	[1.25, 1.60]	<0.001
Musculoskeletal/integumentary diseases
Connective tissue disease	1.44	[1.23, 1.68]	<0.001
HIV infection	2.05	[1.44, 2.92]	<0.001

CI: confidence interval; HIV: human immunodeficiency virus.

Adjusted for the variables with p < 0.001 in Tables 1 and 2. Variables regarding medical illnesses with p < 0.001 were retained in the final multiple conditional logistic regression.

Incidence of medical illnesses within 1 year after bipolar disorder diagnosis

We estimated the incidence of each specific medical illness across organ systems in the patient cohort and control cohort. Table 4 presents the incidence and aIRRs of specific medical illnesses between the patient cohort and the control cohort during the first year after the index date. Notably, the aIRRs of the medical illnesses were >1.00 across the organ systems. In particular, HIV infection, cerebrovascular diseases, ischaemic heart disease, congestive heart failure, other forms of heart disease, chronic kidney disease and renal failure had the highest aIRRs within the first year.

Table 4.

Incidence and adjusted incidence rate ratios of medical illnesses in the first year after the index date among patients with newly diagnosed bipolar disorder compared with age- and sex-matched controls.

	Patients(N = 11,884)			Controls(N = 47,536)			aIRR^a	95% CI	P value
	n	Person-years of total cohort	Crude incidence /100,000	n	Person-years of total cohort	Crude incidence /100,000
Cardiovascular diseases
Hypertension	150	11,002	1363.4	282	46,175	610.72	2.23	[1.83, 2.72]	<0.001
Ischaemic heart disease	103	11,247	915.8	123	46,693	263.42	3.48	[2.68, 4.52]	<0.001
Congestive heart failure	34	11,679	291.1	28	47,284	59.22	4.92	[2.98, 8.11]	<0.001
Other forms of heart disease	277	10,026	2762.8	360	44,861	802.48	3.44	[2.94, 4.03]	<0.001
Diseases of arteries, arterioles and capillaries	53	11,489	461.3	89	46,856	189.94	2.43	[1.73, 3.41]	<0.001
Diseases of veins and lymphatics	322	10,126	3179.9	553	44,005	1256.68	2.53	[2.21, 2.90]	<0.001
Cerebrovascular diseases	131	11,238	1165.7	74	46,893	157.81	7.39	[5.55, 9.82]	<0.001
Respiratory diseases
URI	587	5004	11,730.6	2089	22,779	9170.73	1.28	[1.17, 1.40]	<0.001
Pneumonia	239	10,409	2296.1	355	44,588	796.18	2.88	[2.45, 3.40]	<0.001
Asthma	144	10,363	1389.6	336	44,007	763.51	1.82	[1.50, 2.21]	<0.001
Chronic bronchitis	126	10,837	1162.7	287	45,260	634.11	1.83	[1.49, 2.26]	<0.001
COPD	182	8078	2253.0	514	35,262	1457.66	1.55	[1.31, 1.83]	<0.001
Tuberculosis	27	11,630	232.2	42	47,143	89.09	2.61	[1.61, 4.23]	<0.001
Gastrointestinal diseases
Chronic hepatic disease	386	9210	4191.1	664	42,535	1561.07	2.68	[2.37, 3.04]	<0.001
Ulcer	447	7334	6094.9	876	36,192	2420.42	2.52	[2.25, 2.82]	<0.001
Irritable bowel syndrome	189	10,481	1803.3	366	44,991	813.5	2.22	[1.86, 2.64]	<0.001
Renal diseases
Chronic kidney disease and renal failure	35	11,684	299.6	31	47,249	65.61	4.57	[2.82, 7.40]	<0.001
Endocrine diseases
Diabetes mellitus	120	11,147	1076.5	185	46,325	399.35	2.70	[2.14, 3.39]	<0.001
Hyperlipidaemia	232	10,845	2139.2	384	45,605	842.01	2.54	[2.16, 2.99]	<0.001
Musculoskeletal/integumental diseases
Connective tissue disease	80	11,009	726.7	157	45,806	342.75	2.12	[1.62, 2.78]	<0.001
Atopic dermatitis	110	10,987	1001.2	327	44,862	728.9	1.37	[1.11, 1.70]	0.004
HIV infection	16	11,760	136.1	8	47,429	16.87	8.07	[3.45, 18.85]	<0.001
Cancer	49	11,480	426.8	82	46,847	175.04	2.44	[1.71, 3.48]	<0.001

aIRR: adjusted incidence rate ratio; CI: confidence interval; COPD: chronic obstructive pulmonary disease; HIV: human immunodeficiency virus; URI: upper respiratory tract infection.

Estimated using Poisson regression model.

Incidence of medical illnesses across the first 5 years after bipolar disorder diagnosis

Figure 1 illustrates the aIRRs of specific medical illnesses between the patient cohort and control cohort within 5 years after the index date (for detailed information, see e-Tables 2–5 in the Supplemental Material). Notably, the aIRRs of most medical illnesses between patients and controls were consistently greater than 1.00 throughout the first 5 years after the index date. In addition, among patients with bipolar disorder, the 5-year trends of the aIRRs based on the first-year incidence rates as the reference revealed a decline for cardiovascular diseases (e.g. ischaemic heart disease, other forms of heart disease and diseases of veins and lymphatics), cerebrovascular diseases, gastrointestinal diseases (e.g. chronic hepatic disease and ulcer disease) and communicable diseases (e.g. HIV infection, upper respiratory tract infection and pneumonia) (for detailed information, see e-Table 6 in the Supplemental Material).

Subgroup analyses of patients with adolescent-onset bipolar disorder (i.e. age of onset: 13–18 years) showed that the aIRR in the first year of bipolar disorder was the highest for cerebrovascular diseases, followed by connective tissue disease and hypertension (for detailed information, see e-Table 7 in the Supplemental Material). On the contrary, for patients with adult-onset bipolar disorder (i.e. age of onset: >18 years), the aIRR in the first year following bipolar disorder diagnosis was the highest for HIV infection, followed by cerebrovascular diseases and congestive heart failure (for detailed information, see e-Table 8 in the Supplemental Material).

Sensitivity analysis

In the first sensitivity analysis of patients aged between 13 and 29 years, we used ischaemic heart disease and cerebrovascular diseases as examples; the results revealed aIRRs of 2.54 and 7.12 within the first year after bipolar disorder diagnosis, respectively (for detailed information, see e-Table 9 in the Supplemental Material). These findings were similar to those obtained from the main analysis of the entire cohort (3.48 and 7.39). In the second sensitivity analysis that involved a 5-year washout period, the aIRRs of ischaemic heart disease and cerebrovascular diseases within the first year after bipolar disorder diagnosis were 3.94 and 9.43, respectively, and were similar to those obtained from the main analysis of the entire cohort (e-Table 10 in the Supplemental Material).

Discussion

This is the first nationwide population-based cohort study to investigate the incidence and IRRs of medical illnesses in early-stage bipolar disorder. Our data provided detailed information on the trends of the incidence and aIRRs of medical illnesses in the early stages of bipolar disorder, revealing that the aIRRs of most medical illnesses were consistently greater than 1.00 within the first 5 years following bipolar disorder diagnosis. Notably, cerebrovascular diseases, cardiovascular diseases, renal diseases and HIV infection had the highest aIRRs during the first year after bipolar disorder diagnosis. In addition, observing the 5-year trends of the aIRRs revealed a decline in the aIRRs of cardiovascular diseases, cerebrovascular diseases, gastrointestinal diseases and communicable diseases such as HIV infection. Consistent with the findings of previous studies (Crump et al., 2013; Firth et al., 2019; Forty et al., 2014; Leboyer et al., 2012), our results confirm that bipolar disorder is a multisystem disease that begins manifesting in the early stage of illness.

In this study, cerebrovascular diseases and cardiovascular diseases were the noncommunicable medical illnesses with the highest incidence risk during the first year after bipolar disorder diagnosis. Our findings are in line with those of studies that have indicated an increased vascular risk among patients with early-onset bipolar disorder (Goldstein et al., 2015; Hatch et al., 2015; Naiberg et al., 2017; Westman et al., 2013). Although most well-known vascular risk factors are lifestyle-related (Arnett et al., 2019; Kotseva et al., 2019), our results, along with those of previous studies (Goldstein et al., 2015; Hatch et al., 2015; Naiberg et al., 2017; Westman et al., 2013), suggest that vascular risks are similar in patients with bipolar disorder across different countries where people have different lifestyles. These findings imply that in addition to lifestyle-related factors, certain vascular risk factors might be inherent to the disease process of bipolar disorder.

Our analyses also demonstrated that the aIRRs of cerebrovascular and cardiovascular diseases decreased over the 5-year period after the diagnosis of bipolar disorder. Evidence indicates that vascular risk begins to increase in patients with bipolar disorder since adolescence and young adulthood (Goldstein et al., 2015; Hatch et al., 2015; Naiberg et al., 2017). Somatic diseases are often underdiagnosed in this population (Goldstein et al., 2015; Liu et al., 2017). Therefore, our observations on the downward trends of the aIRRs of cerebrovascular disease and cardiovascular disease may reflect the trends of diagnosis at the time of clinical encounters. However, these downward trends could also be attributed to the potential benefits of the psychiatric care for patients with bipolar disorder. For instance, lithium is associated with a reduced risk of cerebrovascular diseases and cardiovascular diseases in patients with bipolar disorder (Lan et al., 2015; Prosser and Fieve, 2016), likely through the inhibition of proinflammatory responses, reactive oxidative stress and endothelial dysfunction (Kim et al., 2015; Lee et al., 2014). Furthermore, adjunctive psychosocial therapy can prevent relapse and enhance mood stability (Miklowitz et al., 2021), which in turn improves adherence to somatic treatment in patients with bipolar disorder (Levin et al., 2017). Few studies have evaluated whether contact with the mental health system or a specific treatment modality offers protective effects against cardiovascular risk factors in the bipolar disorder population. Considering that patients with bipolar disorder are at a high risk of cardiovascular mortality (Chen et al., 2020a; Westman et al., 2013), future studies are urgently required in this area. Such data could guide further trial studies on the effectiveness of a specific intervention or treatment modality against the initiation and progression of cardiovascular diseases in patients with bipolar disorder.

Among communicable diseases, HIV infection had the highest aIRR during the first year after bipolar disorder diagnosis, and its aIRR declined across the first 5 years after diagnosis. Studies have shown that adolescents and young adults with bipolar disorder have an increased risk of sexually transmitted diseases such as HIV infection (Chen et al., 2019a). The mood symptoms of bipolar disorder and the comorbidities of substance and alcohol use disorders increase the risk of sexually transmitted diseases, including HIV infection, among patients with bipolar disorder (Krantz et al., 2018; Meade et al., 2011). In particular, use of mood stabilisers and second-generation antipsychotics attenuates the observed risk (Chen et al., 2019a). The protective effects of psychopharmacotherapy against sexually transmitted diseases may be mediated by the reduction in affective symptoms, comorbid substance and alcohol use disorders, and risky behaviours (Chen et al., 2019a), which may also explain our observation.

Notably, the aIRRs of renal diseases were consistently >3.50 across the 5-year period following the diagnosis of bipolar disorder. The underlying mechanisms remain unclear and should be investigated in further research. A possible explanation is related to the nephrotoxicity of psychotropic medications used to treat bipolar disorder. Lithium and anticonvulsant mood stabilisers increase the risk of chronic kidney disease by approximately twofold in patients with bipolar disorder (Kessing et al., 2015). Furthermore, lithium may increase the risk of renal failure in patients with coexisting somatic diseases (Aiff et al., 2019). Nevertheless, other studies have not observed this association between lithium use and the risk of chronic kidney disease or renal failure (Kessing et al., 2017; McKnight et al., 2012). Therefore, future studies should investigate whether exposure to psychotropic medications contributes to the increased risk of renal diseases in early-stage bipolar disorder.

In this study, although the aIRRs of certain medical illnesses were the highest in the first year and reduced steadily in each subsequent year following bipolar disorder diagnosis, other medical illnesses did not exhibit this trend. Although we proposed some possible explanations for these complex observations, understanding the nature of these complex manifestations requires further research. A series of studies need to target each specific medical illness and identify factors that aggravate or alleviate the initiation and progression of each in the bipolar disorder population. Furthermore, studies must also integrate medical illnesses as an extended dimension in the multidimensional staging model and examine the correlations of clinical courses of medical illnesses and other core components of bipolar disorder. Such data could be used in the development of a clinical staging model of multisystem bipolar disorder and meaningfully guide mental health workers in the delivery of multidisciplinary healthcare.

Studies have increasingly suggested that adolescent- and adult-onset bipolar disorder may be distinct subtypes in terms of their pathophysiology, psychiatric comorbidity and clinical outcomes (Baldessarini et al., 2012; Baykara et al., 2021; Frahm Laursen et al., 2019; Goldstein and Levitt, 2006). However, few studies have investigated the initiation and progression of medical illnesses between these two subtypes during the early stage of bipolar disorder. Our subgroup analyses revealed that patients with adolescent-onset bipolar disorder tended to be at an increased risk of connective tissue disease in addition to cerebrovascular and cardiovascular diseases in the first year of bipolar disorder, but those with adult-onset bipolar disorder were at an elevated risk of HIV infection in addition to cerebrovascular and cardiovascular diseases. The results indicated that patients with adolescent- or adult-onset bipolar disorder may have differential risks of certain medical illnesses during the progression of bipolar disorder. Given the relatively small sample size of patients with adolescent-onset bipolar disorder in this study, larger naturalistic clinical cohorts are required to examine the risk of various medical illnesses between adolescent- and adult-onset bipolar disorder.

Strengths and limitations

The use of a Taiwanese population–based database enabled us to investigate the incidence of medical illnesses in a representative bipolar disorder cohort. Therefore, our results are less confounded by selection bias. Moreover, the NHIRD comprises detailed information on the diagnostic codes of all NHI beneficiaries, allowing us to examine various types of medical illnesses across different organ systems.

Nevertheless, our study has several limitations. First, in this study, we matched patients and controls based on age and sex to limit potential confounding by these two factors. However, statistical power may have been reduced, limiting our ability to detect differences in certain medical illnesses between patients and controls. In addition, this study calculated incidence rates using person-exposure years, which could be problematic when hazards (i.e. the risk of adverse events) are not constant over time (Kraemer, 2009). However, we used adjusted IRR as the index based on Poisson regression model for the best estimation. Second, the diagnoses of psychiatric and medical illnesses were based solely on ICD-9-CM codes, and hence, their validity could be questioned. However, the NHI Administration randomly reviews medical records to verify diagnoses each year; consequently, the NHIRD has acceptable coding accuracy for epidemiological analysis. Third, although we applied a washout period from 1 January 1996 to 31 December 1997 to exclude prevalent cases, some cases in our study may have been prevalent because bipolar disorder can occur much earlier during teenage years and early adulthood (Goldstein et al., 2017). This concern was especially relevant to 38% of the cohort aged between 30 and 40 years. Our sensitivity analyses conducted using a subgroup aged between 13 and 29 years as well as another subgroup that entered the cohort between 2001 and 2011 (i.e. allowing for a 5-year washout period) also revealed increased IRRs of medical illnesses across all organ systems. However, we were unable to dismiss the possibility of an index mood episode or preexisting medical illness prior to 1 January 1996. Similarly, the order of illness onset based on the diagnosis at the time of clinical encounters may not reflect the true directional links between medical and mental disorders before 1 January 1996. Fourth, to improve diagnostic accuracy, we selected inpatients with bipolar disorder from the PIMC database. Therefore, our findings may not be generalisable to patients without records of psychiatric admission. Fifth, because of limitations inherent to the NHIRD, we could not obtain variables related to lifestyle or compliance with medical treatment. In addition, we were unable to collect laboratory test data from the NHIRD. These data may be helpful for future research on pathogenetic mechanisms underlying the association between medical illnesses and bipolar disorder. Sixth, the mortality rates of patients with bipolar disorder were higher than those of the controls during the 5-year period after the index date, especially within the first-year period (data not shown). Survivorship bias may have affected our analyses of the IRR, particularly for medical illnesses with a high risk of mortality at a young age.

Conclusion

Despite the aforementioned limitations, this national population–based cohort study revealed that patients with bipolar disorder have an increased risk of numerous medical illnesses across multiple organ systems in the early stages of illness, particularly cerebrovascular diseases, cardiovascular diseases, renal diseases and communicable diseases such as HIV infection. In addition, their incidence risks, except for those of renal diseases, declined across the 5-year period after the diagnosis of bipolar disorder. To obtain a deeper understanding of the progression of medical illnesses in patients with bipolar disorder during the early stages, additional studies are required to investigate both risk and protective factors specific to individual medical illnesses in this at-risk population.

Supplemental Material

sj-doc-1-anp-10.1177_00048674211046891 – Supplemental material for Prevalence and 5-year trend of incidence for medical illnesses after the diagnosis of bipolar disorder: A nationwide cohort study

Supplemental material, sj-doc-1-anp-10.1177_00048674211046891 for Prevalence and 5-year trend of incidence for medical illnesses after the diagnosis of bipolar disorder: A nationwide cohort study by Pao-Huan Chen, Shang-Ying Tsai, Chun-Hung Pan, Yi-Lung Chen, Sheng-Siang Su, Chiao-Chicy Chen and Chian-Jue Kuo in Australian & New Zealand Journal of Psychiatry

Footnotes

Acknowledgements

This manuscript was edited by Wallace Academic Editing.

Author Contributions

Drs P.-H.C. and C.-J.K. conceived and designed the study. Dr C.-J.K. acquired the data. Mr S.-S.S. performed the statistical analysis. Drs P.-H.C. and C.-J.K. drafted the manuscript. Drs C.-H.P. and Y.-L.C. made critical revisions to the manuscript for intellectual content. Drs S.-Y.T. and C.-C.C. provided administrative and material support and supervised the study.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: This research was supported by grants from the Ministry of Science and Technology, Taiwan (MOST 105-2314-B-532-006-MY3 and MOST 108-2314-B-532-005) and Taipei City Hospital (10601-62-002). The funding sources had no involvement in the study design, data collection, data analysis, data interpretation, writing of the report or the decision to submit the paper for publication.

ORCID iDs

Shang-Ying Tsai

Chian-Jue Kuo

Data Availability Statement

Please use the following website () to request information on the database used in this study, such as the point of contact, data protection and content of data files. The database is composed of medical claims files representative of the entire population in Taiwan.

Supplemental Material

Supplemental material for this article is available online.

References

Aiff

Attman

Ramsauer

, et al. (2019) Cardiovascular comorbidity increases the risk for renal failure during prophylactic lithium treatment. Journal of Affective Disorders 243: 416–420.

Arnett

Blumenthal

Albert

, et al. (2019) 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease: Executive summary: A report of the American College of Cardiology/American Heart Association Task Force on clinical practice guidelines. Journal of the American College of Cardiology 74: 1376–1414.

Baldessarini

Tondo

Vazquez

, et al. (2012) Age at onset versus family history and clinical outcomes in 1,665 international bipolar-I disorder patients. World Psychiatry 11: 40–46.

Baune

Malhi

(2015) A review on the impact of cognitive dysfunction on social, occupational, and general functional outcomes in bipolar disorder. Bipolar Disorders 17: 41–55.

Baykara

Koc

Resmi

, et al. (2021) Brain-derived neurotrophic factor in bipolar disorder: Associations with age at onset and illness duration. Progress in Neuro-Psychopharmacology & Biological Psychiatry 108: 110075.

Berk

Post

Ratheesh

, et al. (2017) Staging in bipolar disorder: From theoretical framework to clinical utility.. World Psychiatry 16: 236–244.

Charlson

Pompei

Ales

, et al. (1987) A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. Journal of Chronic Diseases 40: 373–383.

Chen

Wei

Bai

, et al. (2019a) Sexually transmitted infection among adolescents and young adults with bipolar disorder: A nationwide longitudinal study. Journal of Clinical Psychiatry 80: 18m12199.

Chen

Tsai

Pan

, et al. (2019b) Mood stabilisers and risk of stroke in bipolar disorder. British Journal of Psychiatry 215: 409–414.

10.

Chen

Tsai

Pan

, et al. (2020a) Incidence and risk factors of sudden cardiac death in bipolar disorder across the lifespan. Journal of Affective Disorders 274: 210–217.

11.

Chen

Pan

Chang

, et al. (2020b) Physical illnesses before diagnosed as schizophrenia: A nationwide case-control study. Schizophrenia Bulletin 46: 785–794.

12.

Correll

Solmi

Veronese

, et al. (2017) Prevalence, incidence and mortality from cardiovascular disease in patients with pooled and specific severe mental illness: A large-scale meta-analysis of 3,211,768 patients and 113,383,368 controls. World Psychiatry 16: 163–180.

13.

Crump

Sundquist

Winkleby

, et al. (2013) Comorbidities and mortality in bipolar disorder: A Swedish national cohort study. JAMA Psychiatry 70: 931–939.

14.

Evans-Lacko

Zeber

Gonzalez

, et al. (2009) Medical comorbidity among youth diagnosed with bipolar disorder in the United States. The Journal of Clinical Psychiatry 70: 1461–1466.

15.

Firth

Siddiqi

Koyanagi

, et al. (2019) The Lancet Psychiatry Commission: A blueprint for protecting physical health in people with mental illness. Lancet Psychiatry 6: 675–712.

16.

Forty

Ulanova

Jones

, et al. (2014) Comorbid medical illness in bipolar disorder. British Journal of Psychiatry 205: 465–472.

17.

Frahm Laursen

Valentin

Licht

, et al. (2019) Longitudinal outcomes in pediatric- and adult-onset bipolar patients compared to healthy and schizophrenia controls. Bipolar Disorders 21: 514–524.

18.

Goldstein

Levitt

(2006) Further evidence for a developmental subtype of bipolar disorder defined by age at onset: Results from the national epidemiologic survey on alcohol and related conditions. American Journal of Psychiatry 163: 1633–1636.

19.

Goldstein

Birmaher

Carlson

, et al. (2017) The International Society for Bipolar Disorders Task Force report on pediatric bipolar disorder: Knowledge to date and directions for future research. Bipolar Disorders 19: 524–543.

20.

Goldstein

Carnethon

Matthews

, et al. (2015) Major depressive disorder and bipolar disorder predispose youth to accelerated atherosclerosis and early cardiovascular disease: A scientific statement from the American Heart Association. Circulation 132: 965–986.

21.

Grande

Berk

Birmaher

, et al. (2016) Bipolar disorder. Lancet 387: 1561–1572.

22.

Hatch

Andreazza

Olowoyeye

, et al. (2015) Cardiovascular and psychiatric characteristics associated with oxidative stress markers among adolescents with bipolar disorder. Journal of Psychosomatic Research 79: 222–227.

23.

Hung

Yang

Huang

, et al. (2014) Cancer incidence in people with affective disorder: Nationwide cohort study in Taiwan, 1997–2010. British Journal of Psychiatry 205: 183–188.

24.

Jauhar

Ratheesh

Davey

, et al. (2019) The case for improved care and provision of treatment for people with first-episode mania. Lancet Psychiatry 6: 869–876.

25.

Jerrell

McIntyre

Tripathi

(2010) A cohort study of the prevalence and impact of comorbid medical conditions in pediatric bipolar disorder. The Journal of Clinical Psychiatry 71: 1518–1525.

26.

Kemp

Gao

Chan

, et al. (2010) Medical comorbidity in bipolar disorder: Relationship between illnesses of the endocrine/metabolic system and treatment outcome. Bipolar Disorders 12: 404–413.

27.

Kemp

Sylvia

Calabrese

, et al. (2014) General medical burden in bipolar disorder: Findings from the LiTMUS comparative effectiveness trial. Acta Psychiatrica Scandinavica 129: 24–34.

28.

Kessing

Feldt-Rasmussen

Andersen

, et al. (2017) Continuation of lithium after a diagnosis of chronic kidney disease. Acta Psychiatrica Scandinavica 136: 615–622.

29.

Kessing

Gerds

Feldt-Rasmussen

, et al. (2015) Use of lithium and anticonvulsants and the rate of chronic kidney disease: A nationwide population-based study. JAMA Psychiatry 72: 1182–1191.

30.

Kim

Bong

Kim

, et al. (2015) Lithium chloride suppresses LPS-mediated matrix metalloproteinase-9 expression in macrophages through phosphorylation of GSK-3beta. Cell Biology International 39: 177–184.

31.

Kotseva

De Backer

De Bacquer

, et al. (2019) Lifestyle and impact on cardiovascular risk factor control in coronary patients across 27 countries: Results from the European Society of Cardiology ESC-EORP EUROASPIRE V registry. European Journal of Preventive Cardiology 26: 824–835.

32.

Kraemer

(2009) Events per person-time (incidence rate): A misleading statistic? Statistics in Medicine 28: 1028–1039.

33.

Krantz

Goldstein

Rooks

, et al. (2018) Sexual risk behavior among youth with bipolar disorder: Identifying demographic and clinical risk factors. Journal of the American Academy of Child and Adolescent Psychiatry 57: 118–124.

34.

Lan

Liu

Lin

, et al. (2015) A reduced risk of stroke with lithium exposure in bipolar disorder: A population-based retrospective cohort study. Bipolar Disorders 17: 705–714.

35.

Leboyer

Soreca

Scott

, et al. (2012) Can bipolar disorder be viewed as a multi-system inflammatory disease? Journal of Affective Disorders 141: 1–10.

36.

Lee

Lin

Chang

(2014) Antiarrhythmic effect of lithium in rats after myocardial infarction by activation of Nrf2/HO-1 signaling. Free Radical Biology & Medicine 77: 71–81.

37.

Levin

Krivenko

Bukach

, et al. (2017) A re-examination of non-psychiatric medication adherence in individuals with bipolar disorder and medical comorbidities. Journal of Nervous and Mental Disease 205: 182–187.

38.

Liu

Daumit

Dua

, et al. (2017) Excess mortality in persons with severe mental disorders: A multilevel intervention framework and priorities for clinical practice, policy and research agendas. World Psychiatry 16: 30–40.

39.

McKnight

Adida

Budge

, et al. (2012) Lithium toxicity profile: A systematic review and meta-analysis. Lancet 379: 721–728.

40.

Meade

Fitzmaurice

Sanchez

, et al. (2011) The relationship of manic episodes and drug abuse to sexual risk behavior in patients with co-occurring bipolar and substance use disorders: A 15-month prospective analysis. AIDS and Behavior 15: 1829–1833.

41.

Miklowitz

Efthimiou

Furukawa

, et al. (2021) Adjunctive psychotherapy for bipolar disorder: A systematic review and component network meta-analysis. JAMA Psychiatry 78: 141–150.

42.

Morris

Puri

Walker

, et al. (2019) Shared pathways for neuroprogression and somatoprogression in neuropsychiatric disorders. Neuroscience and Biobehavioral Reviews 107: 862–882.

43.

Naiberg

Hatch

Selkirk

, et al. (2017) Retinal photography: A window into the cardiovascular-brain link in adolescent bipolar disorder. Journal of Affective Disorders 218: 227–237.

44.

Plana-Ripoll

Pedersen

Agerbo

, et al. (2019) A comprehensive analysis of mortality-related health metrics associated with mental disorders: A nationwide, register-based cohort study. Lancet 394: 1827–1835.

45.

Prosser

Fieve

(2016) Patients receiving lithium therapy have a reduced prevalence of neurological and cardiovascular disorders. Progress in Neuro-Psychopharmacology & Biological Psychiatry 71: 39–44.

46.

Romano

Roos

Jollis

(1993) Adapting a clinical comorbidity index for use with ICD-9-CM administrative data: Differing perspectives. Journal of Clinical Epidemiology 46: 1075–1079; discussion 1081–1090.

47.

Sajatovic

Strejilevich

Gildengers

, et al. (2015) A report on older-age bipolar disorder from the International Society for Bipolar Disorders Task Force. Bipolar Disorders 17: 689–704.

48.

Shah

Scott

McGorry

, et al. (2020) Transdiagnostic clinical staging in youth mental health: A first international consensus statement. World Psychiatry 19: 233–242.

49.

Szklo

Nieto

(2019) Epidemiology: Beyond the Basics, 4th Edition. Burlington, MA: Jones & Bartlett Learning.

50.

Westman

Hallgren

Wahlbeck

, et al. (2013) Cardiovascular mortality in bipolar disorder: A population-based cohort study in Sweden. BMJ Open 3: e002373.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.78 MB