Physical Illnesses Before Diagnosed as Schizophrenia: A Nationwide Case-Control Study

Abstract

Some physical illnesses are potentially associated with the development of schizophrenia. However, few studies have investigated these associations. Here, we examined physical illnesses and medical utilization patterns existing before patients received a diagnosis of schizophrenia. We enrolled a large representative cohort of the general population in Taiwan (N = 1 000 000) and identified 1969 young patients with a new diagnosis of schizophrenia from January 1, 2000 to December 31, 2013. We conducted a nested case-control study based on risk-set sampling. Each case was age-matched and sex-matched with 4 controls selected from the general population. The case and control groups were compared on the basis of various clinical characteristics. Conditional logistic regression was used to estimate the magnitude of risk associated with newly diagnosed schizophrenia. Within the 1 year before the schizophrenia diagnosis, the cases were most likely to visit the psychiatry department, followed by internal medicine and family medicine departments. According to multivariate analysis, compared with the controls, the cases had substantially higher risk of physical conditions in the prodromal phase, including hypertension (adjusted risk ratio [aRR] = 1.93, P = .001), other forms of heart disease (aRR = 2.07, P < .001), cerebrovascular diseases (aRR = 2.96, P = .001), chronic obstructive pulmonary disease (aRR = 1.50, P = .005), asthma (aRR = 1.76, P = .003), and irritable bowel syndrome (aRR = 2.00, P < .001). A wide range of psychiatric diseases and concomitant use of medications were significantly associated with schizophrenia development. In conclusion, several physical illnesses were identified to be associated with schizophrenia development, indicating that people with these illnesses could be vulnerable to schizophrenia.

Introduction

Schizophrenia is a severe mental illness with a worldwide lifetime prevalence of approximately 1%, and the typical onset age is younger than 40 years.^1,2 It is significantly associated with several adverse outcomes during its progression, such as high premature mortality, relatively high prevalence of physical comorbidities, and functional and cognitive decline.^3,4 Additionally, accumulating evidence suggests that comorbid psychiatric illnesses occur in more than 50% of patients with schizophrenia.^5,6 Physical or psychiatric comorbidity is related to global functional impairment and greater severity of psychopathology and may influence help-seeking behaviors.⁷ However, physical or psychiatric comorbidity could appear before the development of schizophrenia.

The prodromal stage of schizophrenia is characterized by nonspecific clinical psychiatric symptoms, such as depressive and anxiety disorders.^8,9 However, somatic symptoms or physical illnesses may be linked to the onset or risk of schizophrenia.¹⁰ From, or even prior to, the onset of schizophrenia, patients frequently meet the criteria for diagnostic physical comorbidities, including cardiometabolic abnormalities, type 2 diabetes, endocrinal alterations, brain injury, epilepsy, circulatory system diseases, immunological diseases, and asthma.^2,10–16 Recently, numerous meta-analyses of epidemiological studies have revealed potential etiological factors that may explain these associations and contribute to the pathogenesis of schizophrenia, such as dysregulation of the cardiometabolic system, immune system, and central nervous system.^11,17–19 These findings suggest that in the prodromal phase of schizophrenia, multiple body systems might be affected. However, data on the association between somatic diseases and schizophrenia development are inconsistent. To address these gaps, evidence-based research must be conducted to discover the strength of association between system abnormalities and schizophrenia development.

We conducted a nested case-control study by using a nationwide population-based insurance claims dataset and investigated the medical utilization patterns of patients within the 1 year before they received a schizophrenia diagnosis. This study design helped us select appropriate controls for comparison, thus avoiding selection bias. Additionally, we explored the effect of physical comorbidities on the development of schizophrenia.

Methods

Data Sources and the Study Cohort

We used Taiwan’s National Health Insurance Research Database (NHIRD), covering the period from January 1, 1996 to December 31, 2013. Implementation of the National Health Insurance (NHI) program in Taiwan began in 1995, and this program covers 99% of the country’s population of 23 million. The NHIRD comprises comprehensive information about registration files and medical claims data for each insured individual, including prescription details and diagnostic codes according to the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM).¹⁷

We used data from the Longitudinal Health Insurance Database 2005 (LHID 2005), which contains all the original claims data of 1 million beneficiaries enrolled in 2005, with these individuals randomly sampled from the 2005 Registry for Beneficiaries (ID) of the NHIRD. Approximately 25.68 million individuals are listed in this registry, with no significant differences between the sex distributions (χ ² = 0.008, df = 1, P = .93) for individuals in the LHID 2005 vs those in the NHIRD. All the registration and claims data of these 1 million individuals, collected by the NHI program, constitute the LHID 2005 (figure 1).

Fig. 1.

Study flow diagram.

Data useful for the identification of beneficiaries and medical care providers are scrambled by the NHI Administration. This study was approved by the Research Ethics Committee on Human Subjects of Taipei City Hospital, Taipei, Taiwan.

Identifying Patients With Newly Diagnosed Schizophrenia

In the LHID 2005 cohort, we identified young patients who received a new diagnosis of schizophrenia (ICD-9-CM code: 295.xx) between January 1, 2000 and December 31, 2013 (figure 1). We excluded individuals who received a schizophrenia diagnosis (295.xx) between 1996 and 1999 (n = 3780) and those with records of mood disorder (ICD-9-CM code: 296.xx) between 1996 and 2013 (n = 40  887). Finally, 3616 patients with schizophrenia remained.

Additionally, we restricted the age range of the study patients to 10–40 years, which is the most common age range for the onset of schizophrenia.^1,2 Eventually, we identified 1969 young patients with newly diagnosed schizophrenia (figure 1).

Nested Case-Control Study

We conducted a nested case-control study on the basis of risk-set sampling. For each of the 1969 patients, we randomly selected 4 age-matched and sex-matched controls from the study cohort that represented the general population of Taiwan (7876 controls). The date of the new schizophrenia diagnosis was defined as the index date. Controls were assigned the same index date as their corresponding case. Patients who were identified during the follow-up were suitable for serving as controls for earlier cases. Furthermore, to confirm that the control patients were alive before the corresponding index date, each patient in the control group was required to have at least 1 claim record after the corresponding index date.

Potential Risk Factors and Variables for Medical Utilization

We defined demographic variables and clinical characteristics, including sex, age, employment, urbanization, and Charlson comorbidity index (CCI) score. The CCI was used to measure the severity of physical comorbidity.²⁰ Urbanization was divided into 5 levels ranging from 1 (indicating the most urbanized) to 5 (indicating the least urbanized).²¹

We investigated various physical illnesses that occurred within the 1 year prior to the index (corresponding) date. Physical illnesses were categorized as cardiovascular disease (CVD), cerebrovascular disease, respiratory diseases, gastrointestinal diseases, endocrinal diseases, and others (table 3). We also investigated psychiatric illnesses, medical service utilization, and concomitant medication use during the 1 year before the (corresponding) index date. The Anatomical Therapeutic Chemical classification system was used to obtain detailed information regarding concomitant medications.²²

Table 3.

Patterns of Physical Comorbidity Within the 1 Year Prior to the Index Date Among Patients With Newly Diagnosed Schizophrenia and Controls (Ratio 1:4), From 2000 to 2013

	Cases (N = 1969)	Controls (N = 7876)
Characteristics	N (%)	N (%)	Unadjusted Risk Ratioa	95% CI	P value
Cardiovascular diseases
Hypertension	50 (2.5)	87 (1.1)	2.39	1.67–3.41	<.001
Ischemic heart disease	20 (1.0)	30 (0.4)	2.74	1.54–4.88	<.001
Other forms of heart disease	50 (2.5)	84 (1.1)	2.42	1.70–3.45	<.001
Congestive heart failure	5 (0.3)	8 (0.1)	2.50	0.82–7.64	.108
Cerebrovascular diseases	19 (1.0)	20 (0.3)	3.80	2.03–7.12	<.001
Respiratory diseases
Pneumonia	28 (1.4)	78 (1.0)	1.44	0.93–2.23	.099
COPD	75 (3.8)	181 (2.3)	1.70	1.29–2.25	<.001
Chronic bronchitis	24 (1.2)	63 (0.8)	1.55	0.96–2.50	.074
Asthma	44 (2.2)	89 (1.1)	2.00	1.39–2.89	<.001
Upper respiratory tract infection	350 (17.8)	1342 (17.0)	1.05	0.92–1.20	.434
Gastrointestinal diseases
Chronic hepatic disease	78 (4.0)	233 (3.0)	1.35	1.04–1.76	.023
Ulcer disease	98 (5.0)	256 (3.3)	1.56	1.23–1.98	<.001
Endocrine diseases
Diabetes mellitus	34 (1.7)	66 (0.8)	2.12	1.39–3.24	<.001
Others
Diseases of arteries, arterioles, and capillaries	6 (0.3)	14 (0.2)	1.72	0.66–4.47	.268
Diseases of veins and lymphatics, and other diseases of circulatory system	35 (1.8)	130 (1.7)	1.08	0.74–1.58	.693
Cancer	7 (0.4)	31 (0.4)	0.90	0.40–2.05	.808
Connective tissue disease	13 (0.7)	45 (0.6)	1.16	0.62–2.14	.646
Moderate or severe renal disease	9 (0.5)	11 (0.1)	3.27	1.35–7.88	.008
HIV infection	5 (0.3)	2 (0.0)	10.00	1.94–51.54	.006
Atopic dermatitis and related conditions	30 (1.5)	84 (1.1)	1.43	0.94–2.18	.092
Irritable bowel syndrome	50 (2.5)	96 (1.2)	2.10	1.49–2.96	<.001
Hyperlipidemia	33 (1.7)	129 (1.6)	1.03	0.69–1.52	.903
Chronic kidney disease	7 (0.4)	7 (0.1)	4.00	1.40–11.40	.010

Note: Including physical illnesses with more than 0.3% in cases.

^aEstimated using univariate conditional logistic regression.

Statistical Analysis

We used conditional logistic regression to determine the significance of the differences between cases and controls in terms of demographics and covariates (sex, age, index year, CCI score, employment, hospital level, and the number of visits to specialists). We further estimated the risk of schizophrenia in relation to all potential variables, particularly physical illnesses. The multivariate regression analysis was performed with backward variable selection. Variables that exhibited a powerful association (P < .01) were retained in the final model.

We subsequently employed an explanatory multivariate model to investigate comorbid physical illnesses as potential factors affecting the development of schizophrenia (table 3). We used the stepwise strategy to select variables with P < 0.01 in the final model. The regression analyses were performed using SAS software 9.4 (SAS Institute). P < .01 was considered statistically significant in all analyses.

Furthermore, to assess the robustness of the results, we conducted sensitivity analyses. First, we repeated the multivariate regression analysis by excluding those aged younger than 15 years (n = 59), which left 1910 cases (7640 controls) aged 15–40 years for the subgroup analysis. Second, we repeated the multivariate regression analyses by using 10 controls per case (case/control: 1969/19 690). The strategy of selecting 10 controls per case was the same as that for 4 controls per case based on risk-set sampling from the study cohort.

Results

Demographic and Clinical Characteristics of Cases and Controls

Table 1 presents the characteristics of the patients in the 2 groups. In the case group, the male-to-female sex ratio was 1.45 (59.2% vs 40.8%), and the patient ages were 10–40 years. According to our univariate regression, the case group had higher CCI scores in the year before the index date compared with the control group. Furthermore, the patients in the case group were more likely to be unemployed than the controls (relative risk [RR] = 3.58, P < .001).

Table 1.

Demographic and Clinical Characteristics of Patients With Newly Diagnosed Schizophrenia and Age- and Sex-Matched Controls (Ratio 1:4) Randomly Selected From the General Population at the (corresponding) Index Date

	Cases (N = 1969)	Control (N = 7876)
Characteristics	N (%)	N (%)	Unadjusted Risk Ratioa	95% CI	P value
Sex
Male	1166 (59.2)	4664 (59.2)	Reference	—	—
Female	803 (40.8)	3212 (40.8)	—	—	—
Ageb, mean ± SD (yrs)	27.74 (7.5)	27.74 (7.5)
10–14	59 (3.0)	236 (3.0)	Reference
15–19	296 (15.0)	1184 (15.0)	—	—	—
20–24	337 (17.1)	1348 (17.1)	—	—	—
25–29	412 (20.9)	1648 (20.9)	—	—	—
30–34	411 (20.9)	1644 (20.9)	—	—	—
35–40	454 (23.1)	1816 (23.1)	—	—	—
Charlson comorbidity indexb
0	1724 (87.6)	7287 (92.5)	Reference	—	—
1	178 (9.0)	482 (6.1)	1.58	1.31–1.89	<.001*
≧2	67 (3.4)	107 (1.4)	2.68	1.96–3.66	<.001*
Urbanizationb
Level 1	765 (38.9)	2567 (32.6)	Reference	—	—
Level 2	549 (27.9)	2826 (35.9)	0.65	0.58–0.74	<.001*
Level 3	263 (13.4)	1075 (13.7)	0.82	0.70–0.96	.012
Level 4	219 (11.1)	872 (11.1)	0.84	0.71–1.00	.045
Level 5	173 (8.8)	495 (6.3)	1.17	0.97–1.42	.103
Employmentb
Yes	777 (39.5)	5188 (65.9)	Reference	—	—
No	1192 (60.5)	2688 (34.1)	3.58	3.20–4.01	<.001*
Hospital levelc
Medical center	502 (25.5)	396 (5.0)	Reference	—	—
Regional hospital	888 (45.1)	601 (7.6)	1.15	0.96–1.38	.137
Local hospital	354 (17.9)	646 (8.2)	0.43	0.35–0.53	<.001*
Local clinic	225 (11.5)	6192 (78.6)	0.03	0.02–0.02	<.001*
No visit	—	41 (0.5)	0.00	—	.967

^a Estimated using univariate conditional logistic regression.

^b At the (corresponding) index date.

^c Cases: at the index date. Controls: the visit closest to the index date (within 1 year); if no visit was recorded within 1 year before or after the index date, then it was coded as “no visit.”

*P < .01.

Specialists Who Diagnosed the New Cases of Schizophrenia

We evaluated the areas of specialization of the physicians who made the new diagnoses of schizophrenia; table 2 reveals that the majority of physicians specialized in psychiatry (84.6%), with the next common specialties being the emergency department (3.5%) and neurology (3.2%). In addition, we assessed the number of visits to specialists within the 1 year prior to the index date; compared with the controls, the cases were most likely to visit specialists in psychiatry (mean = 1.8), followed by those in internal medicine (mean = 1.4) and those in family medicine (mean = 1.4).

Table 2.

Patterns of Medical Utilization Within the 1 year Prior to the Index Date Among Patients with Schizophrenia and Controls (Ratio 1:4)

	Cases	Controls
	(N = 1969)	(N = 7876)
Number of Visits on the Specialistsa,b	Mean (SD)	Mean (SD)	Unadjusted Risk Ratioc	95% CI	P value
Total	12.7 (14.5)	9.0 (9.9)	1.03	1.02–1.03	<.001*
Family practice	1.4 (3.9)	0.9 (2.4)	1.06	1.04–1.07	<.001*
Internal medicine	1.6 (3.8)	1.0 (2.7)	1.06	1.04–1.08	<.001*
Surgery	0.4 (1.6)	0.3 (1.2)	1.08	1.04–1.11	<.001*
Pediatrics	0.4 (1.9)	0.4 (1.8)	1.00	0.98–1.03	.746
Gynecology	0.6 (2.3)	0.8 (2.8)	0.96	0.94–0.98	.001*
Orthopedics	0.3 (1.4)	0.2 (1.2)	1.05	1.01–1.09	.011
Otorhinolaryngology	0.8 (2.1)	1.0 (2.3)	0.97	0.94–0.99	.006*
Ophthalmology	0.4 (1.2)	0.4 (1.3)	1.00	0.97–1.04	.874
Dermatology	0.5 (2.0)	0.6 (2.0)	0.98	0.96–1.01	.251
Neurology	0.3 (1.4)	0.1 (0.7)	1.26	1.19–1.33	<.001*
Psychiatrist	1.8 (4.3)	0.1 (0.7)	2.60	2.36–2.88	<.001*
Rehabilitation	0.2 (2.0)	0.1 (1.1)	1.05	1.02–1.08	.003*
Emergent department	0.3 (1.1)	0.1 (0.4)	1.59	1.45–1.75	<.001*
Dentistry	1.0 (2.1)	1.0 (2.0)	0.99	0.97–1.02	.417
Chinese herb medicine	1.4 (4.1)	1.3 (3.6)	1.01	1.00–1.02	.096
Others	1.4 (4.6)	1.0 (3.1)	1.03	1.02–1.04	<.001*

^aMedical utilization: number of visits to specialists within 1 year before the index date, excluding the number of visits on the index date.

^bIncluding departments with mean visit number > 0.2 among the cases; those with mean visit number < 0.2 were merged into “others.”

^cUnivariate conditional logistic regression.

*P < .01.

Physical Illnesses Diagnosed Before Schizophrenia

Univariate analysis revealed that compared with the control group, the cases had more physical illnesses during the 1 year prior to the schizophrenia diagnosis (table 3). We calculated the unadjusted risk ratio of each physical disease before the diagnosis of schizophrenia. Several physical illnesses were significantly associated with incident schizophrenia. However, the case and control groups did not differ in their distribution of upper respiratory tract infection, which was the most common illness.

Multivariate Analysis of Physical Illnesses Associated With Incident Schizophrenia

Multivariate analysis (table 4) showed that the following variables were significantly associated with newly diagnosed schizophrenia: hypertension (RR = 1.93, P < .001), other forms of heart disease (RR = 2.07, P < .001), comorbid cerebrovascular diseases (RR = 2.96, P < .001), chronic obstructive pulmonary disease (COPD; RR = 1.50, P < .001), asthma (RR = 1.76, P < .001), and irritable bowel syndrome (IBS; RR = 2.00, P < .001).

Table 4.

Multivariate Conditional Logistic Regression of Physical Comorbidity Within the 1 year Prior to the Index Date Among Patients with Newly Diagnosed Schizophrenia and Controls (Ratio 1:4), From 2000 to 2013

	Cases (N = 1969)	Controls (N = 7876)
Characteristic N (%)	N (%)	N (%)	Adjusted Risk Ratioa	95% CI	P value
Physical illness
Hypertension	50 (2.5)	87 (1.1)	1.93	1.33–2.82	.001*
Other forms of heart diseaseb	50 (2.5)	84 (1.1)	2.07	1.44–2.98	<.001*
Cerebrovascular disease	19 (1.0)	20 (0.3)	2.96	1.55–5.65	.001*
COPDc	75 (3.8)	181 (2.3)	1.50	1.13–2.00	.005*
Asthma	44 (2.2)	89 (1.1)	1.76	1.21–2.55	.003*
Irritable bowel syndrome	50 (2.5)	96 (1.2)	2.00	1.41–2.84	<.001*

^aVariables regarding physical comorbidity (table 3) were included for multivariate logistic regression; variables with a strongly significant association (P < .01) were retained in the final model.

^bICD-9 codes: 420.xx-429.xx.

^cChronic obstructive pulmonary disease.

*P < .01.

Regarding psychiatric comorbidities, the following variables were significantly associated with newly diagnosed schizophrenia: comorbid intellectual disabilities, organic mental disorder, depressive disorders, adjustment reaction, alcohol-induced mental disorders, drug-induced mental disorders, anxiety states, and sleep disturbance (supplementary e-table 1). Additionally, concomitant use of psychiatric agents such as mood stabilizers, antipsychotics, antidepressants, and benzodiazepines was common during the period before the diagnosis of schizophrenia (supplementary e-table 2).

In the sensitivity analyses, the subgroup analysis in which those aged younger than 15 years were excluded (supplementary e-table 3) revealed almost identical results as the main analysis (table 4), and the same physical illnesses remained in the final model. Moreover, the results (supplementary e-table 4) were similar to the main analysis when using 10 controls per case, except 2 physical illnesses (moderate or severe renal disease and HIV infection) were included. However, these 2 physical illnesses had considerably lower proportions in the case groups (0.5% and 0.3%, respectively).

Discussion

General Approach and Study Strengths

Protecting the physical health of people with severe mental illness, particularly schizophrenia, is crucial, and strategies for this should also emphasize the provision of adequate physical health care at even the earliest stages of mental illnesses.² The present retrospective nationwide population-based study used a nested case-control design as a novel investigational method to explore preexisting physical illnesses associated with the risk of schizophrenia. The major finding of this study is the moderate to strong association between several preexisting physical illnesses (eg, cerebrovascular disease, other forms of heart disease, IBS, hypertension, asthma, and COPD) and the risk of developing schizophrenia. This study provides useful information regarding the context of seeking health care after receiving a diagnosis of schizophrenia and improving current knowledge regarding increased risk in the prodromal state of schizophrenia.

Medical Utilization

To our knowledge, this is the first study to consider the association between the frequency of visits to specialists and new diagnoses of schizophrenia. Before receiving a schizophrenia diagnosis, the case group frequently visited doctors not only specializing in psychiatry but also in several nonpsychiatric departments, such as family medicine, internal medicine, and Chinese herb medicine. The patients in the prodromal phase of psychosis probably visited these other departments of medicine (especially nonpsychiatric departments) because of their physical illnesses or somatic symptoms. These results provide valuable evidence-based information that can aid the development of effective strategies and intervention services to raise physicians’ awareness regarding physical and psychiatric comorbidities indicating the development of schizophrenia.

The current evidence demonstrates that the majority of individuals with schizophrenia have at least one chronic physical condition concomitant with the course of schizophrenia.²³ The range of medical illnesses that immensely distress patients with schizophrenia is extremely wide; such patients tend to experience physical problems at a younger age than the general population.^23,24 Even before the onset of the first psychotic episode, patients with schizophrenia may already have various physical conditions that began manifesting at a younger age.¹⁰

Stroke and Heart Diseases

Stroke and heart diseases were discovered to be associated with the development of schizophrenia, indicating that cardiometabolic dysfunction is present during early psychosis. The literature suggests that immune dysregulation could be a reason for the codevelopment of cardiometabolic syndrome (CMS) and psychosis.² Accordingly, we determined that vascular diseases can cause schizophrenia. Moreover, evidence suggests the involvement of oxidative stress and inflammation in schizophrenia and related psychotic disorders.^25–27 Changes in oxidative stress or inflammation have been attributed to both psychotic illness and CVD, and elevated cytokines may represent an early surrogate marker of increased risk of CVD in schizophrenia.²⁵ Chronic inflammation has been associated with accelerated atherosclerotic plaque formation, insulin resistance, and increased vascular risk, and cytokines created by inflammatory processes are involved in both CMS and schizophrenia development.²⁸

Regarding the mechanisms underlying schizophrenia development, our results indicate that stroke and heart diseases can exist at the onset of schizophrenia. Thus, our results have major clinical implications. One critical implication for clinicians is that these systems must be routinely examined during the assessment for schizophrenia. Because several antipsychotic medications may worsen cardiometabolic parameters, clinicians should particularly focus on selecting antipsychotic treatments that will have minimal adverse metabolic effects.

Hypertension

We demonstrated an association between the risk of hypertension and schizophrenia development. A study has reported such an association.²⁹ By contrast, a recent meta-analysis did not find an association between schizophrenia and hypertension.³⁰ Nevertheless, the age range of participants in the studies included in the meta-analysis was wide, and the temporal association between hypertension and the development of schizophrenia was unclear.³⁰ Furthermore, hypertension may contribute to the development of CVD, including coronary disease, left ventricular hypertrophy, valvular heart disease, and stroke,^31,32 which are potentially associated with schizophrenia.

Asthma and COPD

Our findings reveal that asthma and COPD were significantly associated with a relatively high risk of schizophrenia. Similarly, studies have discovered associations between autoimmune diseases—specifically asthma and COPD—and increased risk of schizophrenia.^12,33 Many epidemiological studies have suggested that autoimmune inflammatory responses play a major role in the pathophysiology of numerous psychiatric illnesses, particularly schizophrenia.³⁴ Moreover, evidence indicates that abnormalities in the balance of cytokines contribute to the abnormal neural structure in schizophrenia. A possible explanation for these associations is that autoimmune inflammatory responses may increase the permeability of the blood-brain barrier and render the brain vulnerable to proinflammatory substances such as cytokines.³⁵ Therefore, immune components such as cytokines may increase the risk of schizophrenia in vulnerable individuals with asthma or COPD. Future studies should explore the complex mechanisms underlying the autoimmune process are involved in the development of schizophrenia.

Irritable Bowel Syndrome

In this study, the proportion of preexisting IBS in the case group was significantly higher than that in the control group. IBS—the most common functional gastrointestinal disorder—has been associated with increased risk of psychiatric disorders, including depressive disorder, anxiety disorder, sleep disorder, and bipolar disorder³⁶; however, few studies have reported the occurrence of schizophrenia after IBS diagnosis.

The prevalence of IBS among patients with schizophrenia was 19% in one study.³⁷ Furthermore, a study demonstrated that diagnosis of certain gastrointestinal diseases, including IBS, led to alterations in the levels of inflammatory cytokines such as interleukin 1, interleukin 6, and tumor necrosis factor alpha.³⁸ Inflammatory cytokines modulate dopaminergic and glutamatergic neurotransmission^39,40 and considerably affect the hypothalamic-pituitary-adrenal (HPA) system.⁴¹ Because neuroinflammation and HPA system dysfunction are implicated in the development of schizophrenia, shared risk factors may exist between IBS and schizophrenia onset.

Psychiatric Comorbidity

We observed a significant association between a wide range of mental illnesses and a greater risk of schizophrenia. This range comprised psychiatric comorbidities found before the diagnosis of schizophrenia (supplementary etable 1). These findings are consistent with those of previous studies on patients in the prodromal phase of schizophrenia, revealing that this phase is marked by nonspecific clinical conditions, such as depression, anxiety symptoms, and sleep disturbance.⁴² Additionally, several studies have supported the notion that comorbid alcohol and other substance use disorders may contribute to the risk of schizophrenia.^4,43 We also found that individuals with intellectual disabilities were at severe risk of schizophrenia; this result is in line with previous clinical reports.⁴⁴

Limitations

The following limitations warrant consideration when interpreting the results of the present study. First, the patients with schizophrenia were identified by physicians with different specialties who may not have received reliable training for the diagnosis of schizophrenia, leading to diagnostic imprecision. For instance, people with psychotic disorders can go on to develop symptom-explaining medical conditions, such as anti-NMDA receptor encephalitis, which could be misdiagnosed as schizophrenia.

Second, the NHIRD does not contain variables that could mediate the effects of physical illnesses on the development of schizophrenia, including blood pressure, lipid profile, and lifestyle variables (eg, diet, physical activity levels, smoking, substance use, and alcohol consumption). These factors could contribute to the development of CVD among young people who are at ultrahigh risk of psychosis.⁴³ Furthermore, increasing evidence shows that lifestyle variables such as poor nutrition, smoking, and physical inactivity are present from the early stages of psychotic illness or even before receiving treatment.^45–47 Future investigations should clarify how these factors mediate the relationship between physical illnesses and the development of schizophrenia. Third, deterioration of quality of life and various types of psychiatric symptoms can occur before the onset of psychosis; however, information regarding quality-of-life measures and psychiatric symptoms is not present in the LHID 2005 dataset.

Fourth, other potential confounders, including social class or socioeconomic status and ethnicity, are unavailable in the dataset; future studies should match cases and controls on these factors. Fifth, although this study established links between schizophrenia and several physical illnesses, the physical conditions of patients with schizophrenia are often underrated in routine clinical practice. However, we discovered no difference between the 2 groups in the distribution of upper respiratory tract infection, which is the most common illness. This also indicates that the case and control groups could have had similar extents of access to healthcare systems.

Implications

The study results reveal that before receiving the diagnosis of schizophrenia, the case group frequently visited doctors specializing not only in psychiatry but also in other specialties, such as family medicine, internal medicine, and Chinese herb medicine. Specifically, we found that several physical illnesses were associated with the development of schizophrenia, including hypertension, other forms of heart disease, CVD, COPD, asthma, and IBS.

Concerning the primary care implications, physicians should be aware of detecting physical comorbidities among patients in the prodromal phase of schizophrenia to effectively manage the disorder. The integration of physical and mental health care is imperative in primary care. Future studies on the identification of convergent etiopathogenic mechanisms should further examine the hypothesis that some preexisting physical diseases are associated with an increased risk of schizophrenia.

Funding

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 (10701-62-023 and 10801-62-004). The funding sources were not involved in the study design, data collection, data analysis, data interpretation, report writing, or the decision to submit the article for publication.

Authors contributions

Dr Y.-L. Chen, Dr Pan, Dr H.-M Chang, and Dr Kuo conceived and designed the study. Dr Kuo acquired the data. Mr Su performed the statistical analyses. Dr Tsai and Dr C.-C. Chen provided administrative and material support. Dr Y.-L. Chen, Dr C.-K. Chang, and Dr Kuo drafted the manuscript. Dr P.-H. Chen, Dr Y.-L. Chen, and Dr Tai made critical revisions to the manuscript for intellectual content. Dr Tsai and Dr C.-C. Chen supervised the study.