Introduction

It has been recognized for many years that there is a much higher prevalence of smoking in schizophrenia patients than in the general population or patients with other mental disorders [1]–[3]. A meta-analysis based on worldwide studies has revealed that, the pooled smoking prevalence rates in Western schizophrenia patients were 76% for males, 50% for females, and 67% for both genders, respectively [3]. Furthermore, schizophrenia patients are not only more likely to smoke, but also smoke more heavily, have longer smoking duration, and have lower rate of smoking cessation than individuals without schizophrenia [4]. Consequently, high rate of smoking may contribute to reduction in life expectancy and excess morbidity and mortality among people with schizophrenia [5]–[8]. In addition, cigarette smoking also has been demonstrated to be associated with more frequent hospital readmissions, increased financial burden, and higher rate of attempted suicide [9]–[12].

Although cultural context contribute to cigarette smoking [13], the majority of the studies on smoking in schizophrenia patients come from Western countries, whereas few reports have come from Asian countries, especially regarding Chinese patients with schizophrenia. In recent years, a limited number of prevalence studies focusing on smoking of schizophrenia patients have been carried out in mainland China [14]–[20]. Compared with most figures reported in Western literature, the smoking prevalence rates of Chinese schizophrenia patients were relatively lower in men (median = 49%, mean = 48.7%) [15], [17], [21], and considerably lower in women (mean = 4.1%) [14], [19]. Although these Chinese studies varied with regard to sampling methods (i.e., consecutive vs. convenience sampling), samples (i.e., inpatient vs. outpatient), and definitions of current smoking (i.e., currently smoke daily vs. currently smoke every day or on some days), the notable discrepancies in the smoking rates between Chinese and Western patients may still indicate the smoking characteristics of schizophrenia patients have ethnic and/or socio-cultural differences. Thus, the findings of Western studies on smoking might not generalize to Chinese patients.

The underlying causal mechanisms of high frequency of smoking in schizophrenia patients are unclear, especially whether smoking is a risk indicator for the development of the illness, a marker of a more severe illness, or an attempt at self-medication [1], [4], [18]. Several studies viewed smoking as an environment risk factor that causes schizophrenia [22], [23]. In a historical–prospective cohort study, Weiser et al. followed 14,248 male adolescents for 4–16 years, and found a statistically significant association between the number of cigarettes smoked before the onset of illness and risk of developing later schizophrenia in smoking adolescents, with heavier smoking being associated with an almost two-fold increase in risk of schizophrenia [24]. A replication study regarding the relationship between smoking initiation and schizophrenia in Chinese population also found that daily smoking was significantly associated with vulnerability to schizophrenia [25]. On the contrary, another longitudinal study over a 27-year follow-up period revealed that smoking was associated with lower risk of subsequently developing schizophrenia and suggested that smoking might be an independent protective factor for schizophrenia [26]. In China, a study by Ma X et al. showed similar findings, in which, the authors found that smoking may delay full symptoms onset of schizophrenia by reducing some cognitive impairment in the illness and thus smoking might be a protective factor [18]. Hence, the evidence for the association between smoking and vulnerability to schizophrenia is inconclusive.

Previous studies on the clinical features of smoking schizophrenia patients also yielded inconsistent findings. Some found that smoking was significantly correlated with younger age, earlier age at onset (AAO) of illness, increased number of hospitalizations, elevated risk of suicide, higher doses of antipsychotic agents, more antipsychotic drug induced side-effects, and more severe psychotic symptoms [2], [9], [12], [18], [27]–[30]. Thus schizophrenia smokers may be more serious cases, supporting the viewpoint of smoking as a marker of more severe illness process. However, the self-medication hypothesis argues that individuals with schizophrenia smoke to alleviate negative symptoms and side effects of anti-psychotic medication [31]. Several studies found schizophrenia patients who were current smokers had significantly fewer negative symptoms [15], [17], [32] and less extrapyramidal side effects (EPS) [15], [17], [33] than non-smoking patients, supporting the self-medication theory [34]. Unfortunately, any one of these theories mentioned above, could not be used to explain all the clinical findings of existing studies, as there were still many exceptions. For example, studies found that there were no statistically significant differences between schizophrenia smokers and non-smokers with regard to the severity of negative symptoms [14], [16], [19], dystonia and dyskinesia [35], [36]. In general, research regarding the clinical characteristics of smoking individuals with schizophrenia has produced mixed results, which may be primarily due to limitations of study design. Many studies had insufficient sample sizes, used convenient samples, lacked standardized assessment tools, failed to fully address confounding (i.e., antipsychotic medication use), restricted study sample to be patients of one hospital, and, in most cases, recruited chronic rather than acute schizophrenia samples, thus only capturing data of part of the schizophrenia population.

Schizophrenia and related psychotic syndromes can be understood as disorders of adaptation to social context[37]. There is substantial evidence that schizophrenia represents the extreme of a psychosis continuum rather than a discrete illness entity, environmental factors such as taking psychoactive drugs during adolescence and young adulthood, have critical roles in the transition process from less severe levels of psychotic disorder phenotypes to schizophrenia [37]–[39]. Twin and adoption studies suggest that inherited genes make a person vulnerable to schizophrenia and then environmental factors act on this vulnerability to trigger the disorder [37], [40]. Therefore, investigating the smoking pattern of schizophrenia smokers in the context of Chinese society, especially age at regular smoking onset and heavy smoking, and the association between psychotic symptoms and smoking, might increase our understanding of the underlying mechanisms of schizophrenia. Moreover, understanding the prevalence of smoking and its clinical correlates in schizophrenia patients is important and helpful for mental health professionals to develop effective strategies to reduce its harmful consequences. Given the inconclusive findings and limitations of current studies, this study aimed to investigate the prevalence of cigarette smoking and its socio-demographic and clinical correlates in a consecutive sample of schizophrenia inpatients using a single antipsychotic agent in Wuhan and Shenzhen, China.

Methods

Results

Characteristics of subjects

Out of 976 schizophrenia inpatients who were screened during the study period, 617 were on antipsychotic mono-therapy treatment and fulfilled the inclusion criteria. Among them, 56 refused, 93 did not complete assessment, and 39 were ineligible for other reasons, leaving 429 (69.5%) patients completed the interview and physical examination. The mean age at the time of this survey was 33.1 years (standard deviation [SD] = 10.2, range = 18∼60), and 66.9% of the subjects were male. The numbers of patients using promethazine, perphenazine, chlorpromazine, sulpiride, paliperidone, clozapine, risperidone, quetiapine, olanzapine, amisulpride, and aripiprazole were 2, 19, 76, 5, 1, 117, 95, 24, 53, 2, and 35, respectively. Therefore, for the antipsychotics used, the second-generation antipsychotic (SGA) mono-therapy was used more often compared to the first-generation antipsychotic (FGA) (77.4% vs. 22.6%) mono-therapy. Other socio-demographic and clinical characteristics of the schizophrenia inpatients are displayed in the first three columns of Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Socio-demographic and clinical characteristics of inpatients with schizophrenia, split by smoking status.

https://doi.org/10.1371/journal.pone.0088478.t001

Smoking patterns of schizophrenia inpatients

Of the 429 patients, 177 (41.3%) were identified as ever smokers, 3 as former smokers (0.7%) and 174 as current smokers (40.6%); 57.5% of male (N = 165) and 6.3% of female patients (N = 9) were current smokers. Among the current smokers, the mean number of cigarettes consumed per day was 16.1 (SD = 9.0, range = 2–40), while 15.2% of them smoked ≤ 9 cigarettes daily, 43.5% of them smoked 10–19 cigarettes, and 41.4% smoked ≥ 20 cigarettes (heavy smoking). For current smokers, the mean age at regular smoking onset was 20.6 (SD = 4.6) years, and the mean AAO for schizophrenia was 24.9 (SD = 7.4) years, with 75.8% starting regular smoking before the onset of schizophrenia. The age of regular smoking start in patients was significantly earlier than their age of illness onset (paired-samples t test, t = 13.38, P<0.001).

The ISPRs of all patients, male and female patients were 1.11(95%CI: 0.95∼1.29), 1.07 (95%CI: 0.91∼1.24) and 4.64(95%CI: 2.12∼8.82), respectively (Table 2). These results mean that only female patients had a statistically higher prevalence of current smoking than the general population, whereas all patients and female patients did not. Compared with smokers of the general population, current smokers with schizophrenia had significantly lower percentage of heavy smokers (41.4% vs. 51.9%, Z = 2.77, P = 0.023). There was no statistically significant difference in age of initiation of regular smoking between current smoking patients and the general population (mean age at smoking initiation in 2010 GATS China  = 21.2 years, one-sample t test, t = 1.72, P = 0.087).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Indirectly standardized prevalence ratios of current smoking in schizophrenia inpatients, calculating using the results of 2010 Global Adult Survey in China as the standard populations.

https://doi.org/10.1371/journal.pone.0088478.t002

Correlates of current smoking

The socio-demographic and clinical characteristics of current and never smokers are shown in the fourth to seventh columns of Table 1. Univariate analysis (the last two columns of Table 1) showed that current smokers were more likely to be male, be in poor marital status (never married, divorced, widowed, or separated), use alcohol, take FGAs, have less pseudoparkinsonism, have less akathisia, be older, have later AAO, have longer duration of illness, receive long-term antipsychotic treatment, be hospitalized more frequently, be scored higher in PANSS, PANSS-P, PANSS-N, and PANSS-G scales. Multivariate logistic regression analysis (Table 3) considering all variables statistically significant in the univariate analysis (with a Backward: Wald χ2 selection method) revealed that male sex, older age, poor marital status, alcohol use, use of FGAs, longer duration of illness, more frequent hospitalizations, and more severe negative symptoms remained independently associated with smoking after adjustment for all other variables.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Multiple logistic regression for correlates of current smoking.

https://doi.org/10.1371/journal.pone.0088478.t003

Discussion

This study demonstrated that the prevalence of current smoking was 40.6% and the ISPR was 1.11 in Chinese schizophrenia inpatients currently receiving a mono-therapy of antipsychotics. However, both the overall and male-specific prevalence rates of current smoking did not differ significantly between patients and the general population. Only female Chinese inpatients with schizophrenia seem to have higher smoking risk than the general population, but the absolute difference (equals to 3.9%) between the two prevalence rates is really too small. Considering the finding that the female-specific prevalence (6.3%) was too low, and the comparable prevalence between the whole sample and the general population, it is difficult to conclude that Chinese inpatients with schizophrenia are more likely to smoke than the general population. This study further demonstrated that current smokers of patients had significantly lower rate of heavy smoking than those of the general population, indicating that Chinese schizophrenia smokers are not heavy cigarette smokers. Although we found the majority of smokers (75.8%) started regular smoking before the disease onset, the mean age of starting to smoke (20.6 years) was similar to that of the general population; these phenomena might not support the viewpoint that schizophrenia patients start regular smoking in greater numbers during adolescence than the normal population in Western countries [63]. Furthermore, the similar overall prevalence of smoking between our patient sample and the general population may indicate that smoking in Chinese schizophrenia patients is an ingrained habit. In Chinese culture, smoking is a deeply rooted tradition, as most of the Chinese people have already been accustomed to offering others cigarettes or smoking cigarettes provided by others as a gesture of hospitality and friendship[64]. The studied psychiatric hospitals also supplied relatively free environments to let their patients smoke. There is a possibility that Chinese schizophrenia patients and the general population had comparable prevalence of smoking before hospitalization. Under the circumstances, individuals with schizophrenia would still keep their smoking habit even after their admission in the hospital, therefore, no statistically significant difference in the overall prevalence of current smoking was found between the two groups in this study. Another finding that almost all of the ever smokers were current smokers in our study sample seems consistent with such hypothesis.

In addition to showing the differences in smoking characteristics between individuals with schizophrenia and the general population, our study also found lower current smoking prevalence, including gender-specific and overall prevalence, of schizophrenia inpatients than those reported previously in Western studies [US (males: 91%, females:70%) [65], Canada (both genders: 63%) [66], Germany (males: 70%) [67], Spain (males: 81%, females: 13%) [68], and Ireland (males: 92% females: 82%) [69]]. These low prevalence rates in our study are somewhat consistent with the findings of previously published Chinese schizophrenia inpatient studies (males: 42.2%∼52.0%, females: 4.5%∼6.3%) [14], [18], [20]. The discrepancies in the smoking rates between Chinese and Western studies could be due to the exclusion of patients with alcohol/heroin dependence (smoking rate in such patients is expected to be high) [70], and the potential selection bias that only patients using a single antipsychotic agent was included in this study. Moreover, in most literature from western countries, the smoking rate of female schizophrenia patients is very close to that of males[3], but in Chinese schizophrenia patients, women's smoking prevalence is obviously lower than men's. The high ratio of male-to-female smoking prevalence rates is in line with that of the normal population in 2010 GATS China survey[57]. The dramatic gender difference reflects the cultural differences between China and western countries. In general, China is still a male dominated society, and smoking is regarded as the patent and privileges of men. Chinese females who smoke are typically viewed as women with either a wild streak or loose morals. Thus the cultural resistance against women smoking in modern China remains very strong[16], this cultural phenomenon should also be applicable to Chinese patients with schizophrenia, which leads to the low smoking rate in female patients compared with western female patients.

When we study the association between smoking and schizophrenia, as measured by OR, it is noteworthy that the estimated ORs have close relationship with the smoking prevalence rates of the general population, because the conclusions regarding whether patients have higher, lower or comparable risk of smoking, highly depending on the control population used. For example, two studies conducted in Turkey and US reported similar smoking rates in their schizophrenia sample (50% and 51%) [71], [72], but the corresponding smoking rates in Turkey and US general population were quite different (43% and 23%), therefore, a non-significant OR of 1.3 for Turkey patients and a statistically significant OR of 3.5 for US patients were obtained [3]. In our study, we not only found a lower prevalence of smoking, but also revealed a comparable smoking prevalence between patients and the general population. Schizophrenia smokers in our sample are less likely to be heavy smokers as Western smokers with schizophrenia [2]. As a result, smoking might not be associated with the vulnerability to schizophrenia in Chinese population. The potential socio-cultural, biological and public anti-smoking policy factors for these differences need to be further explored.

Analyses of correlates of current smoking revealed that, in schizophrenia inpatients, smoking was more prevalent among patients who were male, were older, were in poor marital status, used alcohol, took FGAs, had longer illness duration, had multiple hospitalizations, and had more severe negative symptoms. These findings indicate that the potential etiologies of smoking in schizophrenia are complex and multifactorial. In this study, male sex, older age (below 60 years old), and unmarried status were identified as significantly risk factors for smoking, these findings are in line with previous reports for both the general population [59], [73] and schizophrenia patients [16], [19], [20]. In general, alcohol drinkers usually smoke and vice versa, as both behaviors are strongly influenced by some shared genetic factors [74]. In agreement with the findings of earlier studies [16], [17], smoking was significantly associated with alcohol use in our schizophrenia sample. Several studies have shown a higher prevalence of smoking in schizophrenia patients with lower educational attainment [2], [19]; however, we did not find such association in our sample, which is consistent with three previous studies [12], [16], [17]. This discrepancy might be attributable to the differences in sample characteristic. Also, other risk factors could mask or affect the role of education level.

It has been suggested that individuals with schizophrenia use tobacco to gain relief from EPS induced by antipsychotic medication. Cigarette smoking increases the activity of hepatic CYP 1A2 enzymes, thus decreasing the concentration of antipsychotic drugs [28], [75]. Therefore, patients on FGAs are more likely to smoke as a result of much EPS from FGA treatment. This study partly replicates earlier findings [15], [16], [31], [33], as its univariate analysis found that a higher likelihood of smoking was associated with less EPS (pseudoparkinsonism and akathisia) and use of FGAs. Smoking could increase clearance of antipsychotics, which in turn makes schizophrenia smokers require higher dosages of antipsychotics than non-smokers, but there was no statistically significant difference in the neuroleptic dosage between smokers and never-smokers in the present study. Moreover, more severe negative symptoms were found to be correlated with current smoking. These findings might suggest the undertreatment of inpatients with schizophrenia. Another reasonable explanation for this phenomenon is that there was a significantly higher proportion of FGA use in smokers, but FGAs are not good at reducing the negative symptoms. Hence, the association between more severe negative symptoms and smoking was found. Although nicotine might alleviate negative symptoms of smokers through stimulating release of dopamine and serotonin in the brain [76], these positive effects might have been weakened by the increased clearance of antipsychotics and undertreatment. Other factor correlated with EPS may mask the role of EPS to smoking, so the statistically significant association between less EPS and smoking disappeared in the multiple logistic regression model. The finding of more severe negative symptoms in smokers when compared to never-smokers is similar to the results of several previous studies [29], [30], [77].

After adjusting confounders, smoking was still associated with more hospital admissions and longer illness duration; both findings indicated that our smoking cases have a more severe disease course. Previous studies also support these findings [15], [17], [29], [30], [77]. Presence of smoking has been reported to be associated with increased risk of hospital readmission and longer duration of hospitalization [78]. The increased severity of illness also may be attributed to high prevalence of alcohol use, as alcohol consumption was founded to be associated with increased risk of hospital admissions in cases with physical disease [79]. There is evidence that severe, treatment resistant schizophrenia is associated with smoking in patients with schizophrenia regardless of gender or alcohol use [80]; this is similar to our finding that smoking was related to more severe illness.

The strength of this study was ensuring the analysis on the association of smoking with types and doses of antipsychotic agents in subjects receiving neuroleptic mono-therapy treatment, by eliminating the masked effect of patients receiving multiple antipsychotic medications. However, this study has several limitations that should be noted. First, our study sample consists of inpatients on antipsychotic mono-therapy only and can’t represent all patients with schizophrenia. Therefore, the results may not be applicable to outpatients and inpatients on antipsychotic poly-therapy. In addition, our sample was selected from specialty psychiatric hospitals of two large cities in mainland China. Although the prevalence and correlates identified in our study seem consistent with several findings in previous reports, whether they also apply to other institutions of China is still questionable. So the findings from this study should be generalized with caution. Second, our study was a cross-sectional, observational study in nature, thus the cause-consequence relationships between smoking and risk factors could not be created. Third, the definition of smoking used in the present study is broader, because it defines a current smoker as an ever smoker who currently smokes daily or on some days [81]. On the contrary, most of the previous clinical studies restricted current smokers to be ever smokers who currently smoke daily. Hence, we may overestimate the smoking prevalence of our schizophrenia sample. Fourth, this study used a questionnaire to record the smoking status of schizophrenia patients, which is not as accurate as several laboratory indicators, such as measuring the level of cotinine (a metabolite of nicotine) in the blood samples. Although we had taken some quality control measures to minimum the measurement bias, biases due to memory effect and social desirability response might still be inevitable in the present study. Finally, since the number of identified female smokers was very limited, males accounted for 95% the smokers; on the other hand, the sample size of female patients in our sample is also small. Thus, those identified correlates of smoking may not be applied to female schizophrenia patients and the prevalence estimation of smoking in female patients is unstable. The smoking characteristics of female people with schizophrenia need to be further studied.

Conclusions

In summary, smoking in Chinese male schizophrenia inpatients, who were receiving antipsychotic mono-therapy, is as prevalent as the general population, and the overall prevalence of smoking in Chinese schizophrenia inpatients is lower than most figures reported in Western literature. Smoking had relationship with gender, age, and marital status; Illness severity-related indicators seemed to have close relationship with smoking in this patient population. Smoking may serve as an important clinical marker of severe illness in individuals with schizophrenia. Given the harmful consequences and its association with more severe illness, effective measures to promote smoking cessation in schizophrenia should be implemented in Chinese patients with schizophrenia.

Acknowledgments

The authors thank all the research staff for their team collaboration work and all the patients involved in this study for their cooperation and support.