JPPS Journal

1.       Abbreviations 

CNS        :               Central Nervous System

BMI        :               Body Mass Index

MDC      :               Main Disease Classifications

ICD-10   :               International Statistical Classification of Diseases and Related Health Problems

WHO      :               World Health Organization 

2.       Introduction 

Hiccups are caused mainly by diaphragmatic myoclonus [1], a brief involuntary twitching of the diaphragmatic muscles, together with coordinated contraction of the glottic closure group of muscles. While transient hiccups rarely influence life or disease prognosis directly, persistent hiccups can disturb verbal communication, sleep, eating, and drinking, and in severe cases can result in weight loss, exhaustion, anxiety, and depression [2]. Thus, hiccup control may be clinically important, especially when hiccups arise as a side effect of lifesaving treatment.

Hiccups are classified based on duration as  bout, persistent, or intractable [3]. Alternatively, hiccups are classified based on etiology as psychogenic, organic, or idiopathic [4]. No sex difference in the frequency of common hiccups has been found in healthy subjects, while onset of persistent or intractable hiccups shows male predominance [5]. Lee et al. reported that hiccups of non-Central Nervous System (CNS) origin are more common among males [6]. In our previous study, we investigated demographic and clinical characteristics contributing to hiccups using the large Japanese Adverse Drug Event Report Database [7]. That study suggested that male sex and tall stature are risk factors for hiccups. Chemotherapy-induced hiccups have also been reported, particularly during treatment with cisplatin [8,9]. In addition, dexamethasone often combined with anticancer drugs may increase the risk of hiccups. Lee et al. succeeded in decreasing induction of hiccups by substituting the antiemetic drug dexamethasone with methylprednisolone in the next course of treatment [10]. Thus, it is possible to define high-risk groups based on patient characteristics, but there are few reports on clinical factors influencing hiccup risk, such as specific primary disease and medication. While the clinical features and pathogenesis of hiccups have been partially elucidated, there has been no complete physical examination of the associations between hiccups and clinical variables such as primary disease. To this end, we examined the associations between hiccups and patient demographics, physical characteristics, primary disease, in-hospital events, and treatments received. 

3.       Materials and Methods 

3.1.  Database 

We selected patients hospitalized at Musashino Red Cross Hospital between April 2014 and December 2014. We retrospectively gathered patient physical examination information, clinical history, disease name, and disease classification from the electronic medical chart system and identified those patients with hiccups during the research period. Data on hiccup patients were connected to all other patient data using their ID numbers to construct a data table (Figure 1). We removed duplicated data for patients readmitted during the research period.

We removed duplicated data for all patients readmitted using an ID number linking relevant clinicodemographic information with hiccup occurrence. Patients were classified into adult and non-adult groups for separate analysis. 

3.2.  Extraction of Patient Information 

We investigated patient demographic and physical characteristics, chemotherapy (YES/NO, regimen), surgical treatment (YES/NO), death within 24 hours of admission (YES/NO), disease name according to the Japanese Major Diagnostic Categories (MDC), and International Classification of Disease (ICD-10).

 3.3.  Classification of Age 

We classified all the patients into two groups based on age. Jacques VS et al. reported the association of age with the incidence of hiccups [11]. We focused on the patient’s age and sex and assumed that age was related with hiccups. The incidence of hiccups suddenly increased after 40 years of age (Figure 2). From these results, we classified all patients into two groups: (1) patients group above 40 years old and (2) patients 40 years or younger.

This figure shows the age and number of patients with hiccups.

 3.4.  Hiccups and Patient Information

We compared sex ratios between hiccup and non-hiccup groups for both the patients 40 years or younger and the patients above 40 years old using the Fisher exact test. Mean age, height, weight, and body mass index (BMI = weight in kg divided by height squared) [12] were compared using an independent samples t-test. Proportions of patients receiving chemotherapy (or not), treated surgically (or not), and deceased within 24 hours of hospitalization (or not) were compared using the Fisher exact test. 

3.5.  Hiccups and Disease Classification 

Disease classification was performed at first hospitalization using MDC, consisting of 18 main categories based on the 22 categories of the ICD-10 established by the World Health Organization (WHO) [13]. Secondly, we compiled a cross-tabulation table of the patients above 40 years old based on two classifications, the presence or absence of hiccups and the presence or absence of a specific disease category. Proportions were compared using the Fisher exact test, and odds ratios (ORs) were calculated for each disease category. Similarly, we compiled a cross-tabulation table based on the presence or absence of hiccups and the presence or absence of a specific disease category in patients 40 years or younger, and compared their proportions using the Fisher exact test. 

We also classified primary disease in patients  above 40 years old at hospitalization according to the more detailed ICD-10 and constructed the same cross-tabulation table. Furthermore, we constructed a scatter plot (volcano plot) [14] of the negative log of the P-value (-log₁₀[P]) from the Fisher exact test (y-axis) versus the log of the ROR (ln[OR]) to identify disease categories most strongly influencing the onset of hiccups.

 3.6.  Multivariable Analysis 

We performed multivariable logistic regression analysis to evaluate the importance of each factor for hiccup risk in the patients above 40 years old. The objective variable was hiccups (YES/NO) and explanatory variables were patient physical characteristics, clinical variables, and ICD-10 category that differed significantly in the univariable analysis.

 3.7.  Statistical Analyses 

All analyses were performed using the JMP®Pro12 software (SAS Institute Inc. NC, U.S.A.). Continuous variables are presented as mean (± standard deviation). A P-value < .05 was considered statistically significant. In all individual analyses, patients with missing values were excluded. We estimated the internal correlation using the pair-wise method. When the square of the Spearman rank-order correlation coefficient [ρ²] was greater than 0.9, we concluded that there was an internal correlation. When there was no internal correlation, we treated these items as independent factors. 

4.       Results 

4.1.   Influences of Patient Clinic Demographic Variables and In-Hospital Events On Hiccup Risk

In total, 7603 patients (> 40 years) and 2343 patients (≤ 40 years) were hospitalized at Musashino Red Cross Hospital between April 2014 and December 2014. Hiccups were identified in 256 patients in the 41 years or older group and the 40 years or younger group. Table 1 summarizes patients’ clinic demographic information and the results of univariable analysis for the non-hiccup and hiccup groups.

In the patients above 40 years old, more than 85% of hiccup group patients were male, while there was no difference in sex ratio in the patients 40 years or younger. Univariable analysis of the patients above 40 years old also revealed significant differences between non-hiccup versus hiccup groups in height (159.5 ± 0.1 cm vs. 164.2 ± 0.6 cm, respectively), and BMI (22.2 ± 0.0 vs. 21.2 ± 0.3, respectively), but no difference in age (69.1 ± 0.2 vs. 69.3 ± 0.9, respectively) and weight (56.9 ± 0.2 kg vs. 57.2 ± 0.9 kg, respectively) was observed. Univariable analysis of in-hospital events also indicated that there were significantly greater proportions of the patients  above 40 years old receiving chemotherapy and of the patients above 40 years’ old who died within 24 hours of hospitalization exhibited hiccups. 

In the patients 40 years or younger, there was no difference in sex ratio between the hiccup (15 males and 13 females) and non-hiccup groups (876 males and 1439 females). Univariable analysis revealed a significant difference between the non-hiccup and hiccup groups in height (136.5 ± 0.9 cm vs. 116.7 ± 7.9 cm, respectively), but not age (21.3 ± 0.3 vs. 17.0 ± 2.8 years, respectively), weight (43.7 ± 0.5 kg vs. 35.6 ± 4.8 kg, respectively), or BMI (20.1 ± 0.1 vs. 18.4 ± 0.8, respectively). Univariable analysis also indicated that a significantly greater proportion patient 40 years or younger receiving chemotherapy exhibited hiccups. 

4.2.  Associations Between Specific Disease Categories and Hiccups in The Group Patients ≤ 40 Years

 The table 2a presents the result of univariable analysis for hiccup incidence according to MDC category regarding the patients 40 years or younger. There was a significantly disproportionate number of hiccup patients in the “Diseases & Disorders of Newborns and Congenital Anomaly,” “Blood Forming Organs, Immunologic Disorders,” “Diseases & Disorders of the Female Reproductive Systems and Pregnancy, Childbirth & the Puerperium,” and “Diseases & Disorders of the Newborns & Congenital Anomaly” categories. We did not scrutinize these correlations in greater detail using the ICD-10 owing to the small number of hiccup patients in the group of patients 40 years or younger. 

4.3.   Associations Between Disease Category and Hiccups in Patients > 40 Years of Age 

Univariable analysis indicated significant differences in hiccup rates for the MDC categories “Diseases & Disorders of the Nervous System,” “Diseases & Disorders of Kidney, Urinary Tract & the Male Reproductive System,” “Diseases & Disorders of the Circulatory System,” “Burns, Injuries, Poisonings & Toxic Effects of Drugs,” “Diseases & Disorders of the Musculoskeletal System,” “Diseases & Disorders of the Female Reproductive System and Pregnancy, Childbirth & the Puerperium,” and “Others” (Table 2a). We also compiled a cross-tabulation table of hiccup rates according to the ICD-10 category and calculated P-values using the Fisher exact test as well as associated ORs. From the results of all categories with five or more patients (Table 2-b), we constructed a scatter plot of ORs and searched for significant differences to identify disease categories more strongly associated with hiccups (Fig. 3). The plot revealed particularly strong associations between hiccups in the patients above 40 years old and “Subarachnoid Hemorrhage from Anterior Communicating Artery,” “Malignant Neoplasm: Bronchus or Lung, Unspecified,” and “Intracerebral Hemorrhage in Hemisphere, Subcortical.” All analyses excluded patients with missing values.

4.4.   Independent Factors Influencing Hiccup Risk 

Finally, we performed a multiple logistic regression analysis including all demographic variables that were identified as significant and disease categories for categories with fewer than four 4 patients and P < 0.05 in the univariable analysis (Table 3). Among physical characteristics, low BMI was identified as an independent risk factor for hiccups. In addition, male sex, chemotherapy, and death within 24 hours of hospitalization were independent risk factors. Finally, the brain disease categories “Subarachnoid Hemorrhage,” “Intracerebral Hemorrhage,” and “Tumor of Brain,” as well as multiple malignant tumors of the renal pelvis, extrahepatic bile duct, bladder, and bronchus/lung were significant independent risk factors.

5.       Discussion 

In the current study, risk factors of hiccups were researched with clinical data from the Musashino Red Cross Hospital, a general hospital with many departments. Although the study was performed in the single institution, many cases were accumulated as in multi-center studies. We collected 256 cases of hiccups, surpassing the case-number of our previous study with the Japanese Adverse Drug Event Report Database including 160 cases of hiccups [7]. Furthermore, we were able to handle more clinic data by accessing the electronic medical record system. Thus, this study will be clinically important for it shows the way to control the symptom of hiccups. 

Jacques VS et al. report the association of age with the incidence of hiccups [11]. However, there was no statistical clarification regarding the association between hiccups and age, gender, or disease. We supposed that age was related to the occurrence of hiccups; thus, we classified all patients into two age groups and analyzed the risk factors. We classified all the patients into two groups (patients ≤ 40 years of age and patients > 40 years) because the incidence of hiccups in male patients suddenly increased after 40 years of age (Figure 2). There were some different results between two groups regarding gender and classification of disease. These results suggest that it is the difference in the characteristics of hiccups between the younger group and others, and there is a difference in the mechanism of hiccups between the two groups. There were no reports that focused on gender, age, or classification of disease regarding hiccups. The present report is an important issue due to discussion about the disease-developing mechanism of hiccups. 

The incidence of persistent or intractable hiccups is significantly higher in males than in females [5]. Lee et al. also indicated that hiccups tend to emerge in females at a younger age compared with males [6]. Our results showed that age was not an independent factor for hiccups in the both groups. In contrast to the patients above 40 years old, male sex was not a risk factor in patients  40 years or younger. Most cases of hiccups in the patients 40 years and younger group occurred in low-weight births, disorders of the blood, immune system, and disorders of the nervous system, suggesting that hiccups in younger patients are mainly due to organic disease. This male predominance has provided few clues to the underlying mechanism. It is possible that the stimulation threshold for hiccups is lower in males than in females owing to greater excitability of afferent or efferent nerves in the hiccup reflex arc [6]. Given the absence of male preponderance in patients 40 years or younger, we suggest different underlying mechanisms, such as possible sex differences in the development of respiratory apparatus or nervous system.

In our previous study, we identified tall stature as a risk factor for hiccups [7]. However, input to this database is voluntary so parameters are often missing. Thus, our conclusions were drawn from group data. In contrast, the current study allowed for the investigation of hiccup risk factors using a smaller but more comprehensive hospital electronic medical chart system. Specifically, we could examine the influence of BMI as well as weight and height. In the univariable analysis, height and BMI were related to hiccups (P <0.001), but only BMI was an independent risk factor for hiccups. The significance of height in the univariable analysis may reflect the strong association between height and sex in adults. We suggest that BMI is a suitable physical index for the evaluation of hiccup risk in adults. Furthermore, this relationship between low BMI and hiccup risk may provide clues to the underlying mechanisms. 

Among hospital events, chemotherapy, and death within 24 hours after hospitalization were identified as independent risk factors for hiccups in the patients above 40 years old using multivariable analysis. Several recent reports have documented hiccup risk owing to chemotherapy. Particularly, there are reports of hiccup cases induced by platinum-based drugs such as oxaliplatin and cisplatin [8,9]. A recent report also highlighted the possible involvement of dexamethasone, which is used as adjunct therapy to suppress nausea and allergic reactions [10]. In the present study, chemotherapy was an independent risk factor for hiccups and the strongest risk factor among in-hospital events. However, further investigation is required to determine whether hiccups are solely a response to chemotherapeutic drugs or are related to tumor pathophysiology. 

Death within 24 hours of hospitalization was an independent risk factor for hiccups, although this relationship was not observed for patients within individual disease categories who died within 24 hours. Hiccups may occur at the end stage of cancer and in response to lifesaving measures. Twycross et al. suggested that the main causes of hiccups in terminal-phase patients are relaxation of the stomach, reflection of the stomach-esophagus, diaphragmatic excitement, excitation of the phrenic nerve, poisoning, and central nervous system cancers [15]. In addition, stimulation of the pharynx rear wall by intubation for lifesaving measures may induce hiccups [16]. However, there are no reports on the relationships among hiccups, intubation, and respiration. It is thought that hiccups in the terminal phase reduce QOL by disrupting conversation and sleep. Therefore, the investigation of hiccup treatments for terminal phase patients is urgently required.

 We also analyzed associations between hiccup incidence and specific ICD-10 and Japanese MDC categories. Although limited numbers in some categories precluded analysis, we identified nine disease categories associated with hiccups, with several strongly associated categories according to volcano plot analysis (Figure 3). Multivariable analysis identified subarachnoid hemorrhage, intracerebral hemorrhage in hemisphere, and tumor of brain as strong independent risk factors among patients with CNS diseases. It is known that the hiccups reflex center lies within the medulla oblongata, although the precise neurological mechanism is unclear [16]. We speculate that many brain tumor and hemorrhages in the above 40 years old group affected this brainstem hiccup center. Alternatively, hiccups may be induced by intubation for artificial respiration and gastric tube insertion during emergency treatment of these diseases. While cerebral hemorrhages were a significant risk factor in the present study, we could not identify the lesion sites. Hiccups disturb both treatment and rehabilitation following neurological diseases; therefore, prevention of hiccups may be particularly important for these patients. Further examination of available neuroimages is warranted to assess the relationship between lesion site and hiccup risk.

The volcano plot represents the relationship between hiccups and classification of disease (ICD10). Numbers with points in the figure correspond to the table under the figure. The x-axis represents the odds ratio (OR) in lnOR scale, and the y-axis represents the P value in a–log₁₀P scale. The red line demarcates the significance threshold P = .05. ORs were calculated using cross-tabulation. Different colors indicate the number of hiccup patients (from 5 in blue to 22 in red). Points in the upper right represent disease categories strongly associated with hiccups. 

In addition to brain lesions, widely distributed neoplasms were also independent risk factors for hiccups according to the multivariable analysis (including renal pelvis, extrahepatic bile duct, neoplasms of bladder, and bronchus or lung). It is possible that these lesions damage nerves of the hiccup reflex arc. It was reported that the vagus nerve participates in both afferent and efferent coordination in the hiccup reflex [1,16], and malignant tumors of the aforementioned organs may stimulate the vagus nerve. In addition, the chemotherapy received in-hospital by most cancer patients may induce hiccups independent of tumor-mediated nerve effects. However, we found no internal correlation between specific cancers and chemotherapy; therefore, it appears that both malignant tumors and the medications used for treatment induce hiccups. Further study of the associations between cancer type, chemotherapy regime, and hiccup risk is clearly warranted as hiccups can reduce treatment tolerance and quality of life. Moreover, such analysis may identify treatment regimens with lower risk. 

The present study was based on a retrospective review of an electronic chart system from a single center and had some missing information. In addition, hiccups may not be reported by patients unless regarded as an unbearable symptom. Thus, our conclusions may be confounded by report bias. 

6.       Conclusion 

We investigated risk factors for hiccups including patient age, physical characteristics, clinical variables, primary disease, and in-hospital events. We report for the first time that BMI is an independent risk factor for hiccups in patients above 40 years old. We also confirmed that chemotherapy, certain malignant tumors, and neurovascular diseases are risk factors. These results may enable the identification of high-risk patients and treatment regimens, thereby reducing risk and improving quality of life. Moreover, such information may help clarify the pathomechanisms of hiccups.

Table 1: Patient physical characteristics and in-hospital events

^#: Hiccup and non-Hiccup groups were compared using the Fisher exact test

*: Hiccup and non-Hiccup groups were compared using a t-test

###:***: p < 0.001

##:**: p < 0.01

*: p < 0.05

a

b

Hiccup and non-Hiccup groups were compared using the Fisher exact test

****: p < 0.0001

***: p < 0.001

**: p < 0.01

*: p < 0.05

Table 2: Disease category and hiccups

Table 3: Results of the multivariable analysis indicating independent risk factors for hiccups treated as variables that do not affect each other.

*: indicated a significant odds ratio

1. Kondo T (2005) Hiccup: Physiologic backbone and the point of the treatment. J Clinical Anesthesia 29: 225-232.
2. Wilcock A, Twycross R (1996) Midazolam for intractable hiccup. J Pain Symptom Manage 12: 59-61.
3. Kolodzik PW, Eilers MA (1991) Hiccups (Singultus) review and approach to management. Ann Emerg Med 20: 565-573.
4. Rousseau P (1995) Hiccups. South Med J 88: 175-181
5. Takemoto T (2009) Hiccups-Diagnosis and treatment. Sougou-Rinsyou 58: 7.
6. Lee GW, Kim RB, Go SI, Cho HS, Lee SJ, et al. (2016) Gender differences in hiccup patients: analysis of published case reports and case-control studies. J Pain Symptom Manage 51: 278-283.
7. Hosoya R, Uesawa Y, Ishii-Nozawa R, Kagaya H (2017) Analysis of factors associated with hiccups based on the Japanese Adverse Drug Event Report Database. PLoS One 12: e0172057.
8. Hayashi M, Sugimura H, Suga Y, Kawahara M, Aimiya K, et al. (2009) Study on risk factors for hiccups induced by cisplatin-based chemotherapy. Jpn J Pharm Health Care Sci 35: 89-95.
9. Liaw CC, Wang CH, Chang HK, Wang HM, Huang JS, et al. (2005) Cisplatin-related hiccups: male predominance, induction by dexamethasone, and protection against nausea and vomiting. J Pain Symptom Manage 30: 359-366.
10. Lee GW, Oh SY, Kang MH, Kang JH, Park SH, et al. (2013) Treatment of dexamethasone-induced hiccup in chemotherapy patients by methylprednisolone rotation. Oncologist 18: 1229-1234.
11. Souadjian JV, Cain JC (1968) Intractable Hiccups Etiologic Factors in 220 Cases. Postgraduate Medicine 2: 72-77.
12. Centers for Disease Control and Prevention.
13. ICD-10 Version: 2014.
14. Chen JJ, Wang SJ, Tsai CA, Lin CJ (2007) Selection of differentially expressed genes in microarray data analysis. Pharmacogenomics J 7: 212-220.
15. Twycross R, Wilcock A (2001) Symptom Management in Advanced Cancer. (3^rd Edition). Oxford: Radcliffe Medical Press, USA.
16. Oshima T, Sakamoto M, Arita H (1994) Hiccup like response elicited by mechanical stimulation of dorsal epipharynx. J Appl Physiol 76: 1888-1895.