Association between Appendectomy and Demyelinating Disorders in Subjects 40 Years and Older
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.293469
Source of Support: None, Conflict of Interest: None
Keywords: Appendicitis, epidemiology, microbiome, microflora, multiple sclerosis, neuromyelitis optica
Multiple sclerosis (MS) and neuromyelitis optica spectrum disorder (NMOSD) are immune-mediated disorders of poorly understood pathology., Neurologists wonder if the microbiome has any role to play.,, A review showed that certain bacterial families, specifically Butyricimonas strains, Faecalibacterium, Bacteroidaceae, Ruminococcus, Lactobacillaceae, and Clostridium, were probably more prevalent in MS subjects than in controls. However, the papers in the review showed little consistency in findings. While it is certainly possible that the microbiome is involved in the pathogenesis of MS and NMOSD, it is unlikely that open-ended searches for organisms will yield results since the microbiome (at the time of sampling) may be modified by disease, treatment, and diet.
We can, however, look at the microbiome from a different direction. It is possible that some diseases will occur in clusters if they carry similar patterns. Acute appendicitis may be a microbiome-related condition, and may be considered a starting point for an epidemiologic study, particularly since it has a well-established (negative) association with ulcerative colitis., In fact, this link has prompted researchers to look for associations with several diseases empirically,,,, but reported associations have been weak, and often disproved in subsequent studies.,,
It is unlikely that ulcerative colitis is the only condition associated with appendicitis, so why were researchers unable to find other links? In our opinion, the reason lies in the mistaken hypothesis that the appendectomy itself influences the disease by altering immunity, perhaps by decreasing the bulk of lymphoid tissue or directly through a molecule (such as alfa-synuclein). If one presumes that the appendectomy itself causes, or protects against, an illness, one naturally includes subjects for study only if they undergo appendectomy before the onset of the disease.,
We believe that the appendix itself has no direct effect on the rest of the body. Instead, it is likely that appendicitis and ulcerative colitis share a commonality in their etiopathogenesis. Tsai et al. hint at this commonality. Further, it is appendicitis, not appendectomy, that is associated with ulcerative colitis: removal of a normal appendix does not influence the occurrence of ulcerative colitis. This immediately negates the hypotheses relating to immune modulation and lymphoid bulk. Therefore, when looking for an association between appendicitis and a given disease, we should be searching for a correlation between the lifetime risks of appendicitis and of that disease. Although there is no hard evidence, we strongly believe that the body's microbiome is the key to the etiology of both MS and appendicitis. If there is indeed an association between the two disorders, further studies could identify the common component of the microbiome.
Information derived from younger subjects is unreliable. Those who do not have MS might develop it later. Similarly, those who have not had appendicitis could still develop it later. We, therefore, restricted our data collection to MS and normal individuals aged 40 and older. By this age, most persons who will ever have MS will have developed MS. Also, by the age of 40, most persons who will ever have appendicitis will have had appendicitis. While searching for MS, we added NMOSD, which shares several features with MS and, indeed, has been long considered a variant of MS.
MS and NMOSD patients were recruited from the neurology outpatient clinics at 3 tertiary referral hospitals in West Malaysia: Hospital Kuala Lumpur, Hospital Pulau Pinang, and Hospital Sultanah Aminah (Johor Bahru). These are Malaysia's three largest Ministry of Health centers for the management of MS. Patients were interviewed for a history of appendectomy.
MS and NMOSD subjects
All MS or NMOSD patients were 40 years or older. We extracted the medical records of all patients diagnosed as MS or NMOSD from the hospital databases. The records were then screened by the neurologists (SV, EGB, AC, KR) to confirm that the diagnoses were consistent with the 2010 McDonald's Criteria and Wingerchuk's 2015 Criteria.,
A convenience sampling was carried out to question the controls. A team of trained students gathered information from apparently healthy persons in their vicinity. All controls were also 40 years or older at the time of interview.
The interviews were carried out over the telephone or face-to-face in English, Bahasa Malaysia, or Mandarin according to the patient's preference, and a brief questionnaire was completed. The questionnaire recorded baseline demographic features (age, gender, and ethnicity), and asked about a history of appendectomy. If respondents admitted to an appendectomy, they were further asked if, to their knowledge, the appendix had been inflamed at the time of surgery. The questionnaire also asked if the subjects had experienced any chronic skin, digestive, respiratory, or bone disorder. No interviews were conducted over email or social media. Subjects who knew that they had undergone incidental appendectomy (removal of a normal appendix rather than an inflamed one) were excluded from the study. All other subjects were included.
Sample size was calculated using the following formula:
N = [(Zα+Zβ) 2*(p1(1-p1) + p2(1-p2))]/(p1-p2) 2
We have assigned the values p1 = prevalence of appendectomy in normal subjects = 0.15, and p2 = prevalence of appendectomy in NMOSD or MS subjects, arbitrarily considered to be a third or three times higher than control (similar to the effect size seen for prevalence of appendectomy in ulcerative colitis subjects when compared to controls). We chose a conservative alpha at 0.1 instead of the usual 0.05 because this was an exploratory study (Zα = 1.65 for a type I error of 10%), and set the power at 80% (Zβ = 0.84 for a type II error of 20%). The minimum calculated sample size was 110 for each group (test and control). The test group represented the combined sum of MS and NMOSD subjects since at the beginning of the study we expected the results to be the same for both conditions. This sample size is consistent with that used earlier for determining the association between ulcerative colitis and appendectomy.
The chi-squared test was used for the basic comparison. If the assumption of Chi squared test could not be met, a Fisher's exact test was used instead. Binary logistic regression was conducted to explore the association of multiple variables on the appendectomy rates in normal and MS/NMOSD subjects.
Calculations were carried out on SPSS version 22.
The study was approved by Malaysia's Medical Research Ethics Committee (MREC ID – 33914) on 27 March 2017 and by Monash University's Human Research Ethics Committee (ID – 7938) on 21 April 2017. Good Clinical Practice guidelines were followed.
Overall, 111 MS and 97 NMOSD subjects were screened. Subjects were all 40 years or older. Only 49 MS and 71 NMOSD subjects met the inclusion criteria and were contactable. Overall, 880 controls were also screened and included [Table 1].
Appendectomy rates in MS subjects, NMOSD subjects, and Controls
Of 49 MS subjects, 9 had undergone appendectomy (prevalence rate 0.184, 90% confidence intervals 0.092 to 0.275). The distribution of cases by ethnicity is shown in [Table 2].
Of 71 NMOSD cases, 4 had undergone appendectomy for appendicitis (prevalence 0.056, 90% confidence intervals 0.011 to 0.101). The distribution of cases by ethnicity is shown in [Table 3].
Of 880 controls, 59 had undergone appendectomy for appendicitis (prevalence 0.067, 90% confidence intervals 0.053 to 0.081). The distribution of cases by ethnicity is shown in [Table 4].
We compared the prevalence of appendectomy in the three groups. Comparing MS vs NMOSD, the prevalence rate of appendectomy was significantly higher in MS cases (18.37%) compared to NMOSD cases (5.6%) (Fisher's exact test two-tailed P = 0.036). Comparing MS vs Control, subjects with MS had a higher rate of appendectomy than the control subjects (18.37% vs 6.71%) (Fischer exact test two-tailed P = 0.0068). Finally, comparing NMOSD vs Controls, NMOSD patients and controls had a similar rate of appendectomy (5.8% vs 6.71%, n.s.).
Since the control and MS groups were not perfectly matched in age, gender and ethnicity, we conducted a binary logistic regression analysis to partial out the effects of these three variables. For this, we excluded the NMOSD cases as well as the cases classified for ethnicity as “Others”. We then conducted the regression analysis, setting appendectomy as the dependent variable. [Table 5] shows the SPSS output for the analysis.
The analysis shows that MS is an independent predictor for appendectomy, even when race, gender, and age are accounted for. Persons who have MS have thrice the odds of undergoing appendectomy some time in life (P = 0.009) compared to persons without MS.
In this study, we have looked for an association between appendicitis and two immune-mediated conditions: MS and NMOSD.
Demographics ethnic distribution of MS and NMOSD subjects
Malaysia's ethnic distribution for Malays, Chinese, and Indians is 9.8:3.3:1, respectively. Viswanathanet al, in a study based on 2010 revised McDonald's criteria, found Malays, Chinese, and Indians to constitute 53%, 18%, 27% of their MS subjects respectively. Our findings are consistent with this study. These studies suggest that Indian ethnicity is a risk factor for MS in the Malaysian population. In contrast, among patients diagnosed with NMOSD, there is a high prevalence of ethnic Chinese (43%). It is difficult to compare our findings with other recently published literature,, since the modified criteria for diagnosis of MS and NMOSD were not used consistently in all these reports.
Appendectomy rates in MS, NMOSD, and controls
Appendectomy was much more prevalent in MS patients compared to controls (18.37% vs 6.71%).
This raises two questions. One: is the high prevalence of appendectomy in MS cases a result of the increased numbers of ethnic Indians in this group? Among MS patients, there were more Indians than would be expected, and, as we can see in [Table 4], appendectomy rates are higher among the Indians. Two, why are we showing such a strong association with appendectomy when earlier studies showed a minimally increased odds ratio?
The first question is answered by the binary regression analysis, which shows a high association between appendectomy and MS when the effects of ethnicity are partialed out. MS is an independent predictor of appendectomy, with odds ratios approaching 3.
This brings us to question two: why are our results different from those reported by others? We expect that there are two reasons: one, related to diagnosis, and two, related to the age range that we have analyzed.
Difference in diagnosis
For many published studies, cases diagnosed as MS are likely to have included patients with NMOSD. The current criteria, are new, and have not been applied in several published papers. Lunny et al. reviewed 33 case-control studies involving over 27,000 MS cases and over 200,000 controls. They found a very small increased risk for MS in subjects who had undergone early appendectomy. The studies in their review included some that were published before the current diagnostic criteria were established, therefore, in effect, our study and earlier ones talk about different types of patients.
Difference in age groups
The age of subjects included in our study was forty and above. Previous studies, including newer ones, have made no effort to restrict data collection to subjects over 40. Consequently, some young persons labeled as controls may have developed MS later in life, and several persons labeled as negative for appendectomy may have developed appendicitis in subsequent years.
Why is appendectomy associated with MS?
Our data strongly suggest that appendectomy has a strong positive association with MS.
The mechanism is a matter of conjecture, but we feel that future studies will show a connection through the microbiome. Several experts already believe that MS is a result of an altered microbiome. In animal models of MS, gut-free mice were resistant to the disease. The generation of Th 17 cells, implicated in the pathophysiology of MS, is intimately linked to the gut microbiome., Richardsen et al. believed that bacterial and viral pathogens trigger appendicitis, and Lunny et al. went so far as to say that appendicitis is an autoimmune disease like MS. We hypothesize that the microbiome (probably the gut microflora) patterns the body's immunity to directly influence the risks of appendicitis and of MS. The association between appendicitis and MS is merely indirect [Figure 1]. Other disorders that have been considered for links with appendicitis, are also probably linked indirectly through the microflora.
According to our hypothesis, the association between appendicitis and MS is not direct (A), but, instead, is more likely to be an indirect association due to a common causative factor, probably the microbiome (B).
Our study has a few limitations. Our numbers, particularly of MS, are small. This is because MS is rare in Asia, and even though we attempted to include every case from the three largest government hospitals, we could not achieve a satisfying number, especially since we focused on subjects who were 40 years and older.
Then again, the sampling of controls was not randomized: for this study, it was not possible to gather large numbers of random controls without substantial funds. Nevertheless, we do not believe that a randomized control sample would show a significantly different prevalence of appendectomy.
In patients who underwent appendectomy, we were unable to confirm the histology to ensure that the patient did indeed have acute appendicitis. We attempted to minimize the numbers of incidental appendectomies by excluding patients who knew that their appendices were normal at surgery. Nevertheless, it is quite likely that a few of the patients who reported appendectomy did undergo removal of a normal appendix. Since this would apply to both MS and to control subjects, it is unlikely to alter the final result.
On the other hand, our study has several strengths. We have a very large control size, which allows for adequate comparisons. We have collected MS cases from the three main Ministry of Health centers in Malaysia. Most patients with MS or NMOSD aged over 40 are likely to be registered in one or other of these centers at some time in their lives, and it is likely that we have been able to recruit most eligible cases in the country.
This allowed us to categorize patients into “appendectomy” and “no appendectomy” We believe that it is this inclusion criterion that is the reason for our finding of a strong association between appendicitis and MS. About 10% of persons develop appendicitis after 40, therefore there is still the possibility of a small error in the categories “appendicitis” versus “no appendicitis”. This error could be further reduced by setting the cut-off age at 50 years, but our sample size for MS would have been even smaller.
Our results strongly indicate that MS, but not NMOSD, is associated with a higher prevalence of appendectomy than the normal population. The odds ratio of appendectomy vs no appendectomy in an MS patient is almost three times higher than in a control subject.
The association between MS and appendicitis is unlikely to be direct: it is more likely that the association between MS and appendicitis is indirect, through a common external factor. In our opinion, there are reasons to believe that this external factor is the body's microbiome, and it is important for the microbiome to be carefully studied in order to determine the etiopathogenesis of MS.,
SS conceived the idea for the study. SS, SV, and NPB designed the study. The study was part of a Bachelor of Medical Sciences (BMedSc Hons) research project conducted by the BMedSc Hons student, HSO. SS and NPB were his supervisors. SV, EGB, KR, and CYT reviewed the patient data at Hospitals Kuala Lumpur, Pulau Pinang, and Sultanah Aminah to confirm that the diagnoses of MS and NMOSD conformed to the stated criteria. They also supervised the collection of data at these centers. The data was collected by HSO, SV, EGB, CYT, and IMAR. RBN conducted the statistical analysis. The paper was written by SS, SV, and NPB, and reviewed by all authors.
The following contributed to the collection of data for controls: Kong Jo Yi, Roger Gan Giah Chien, Patricia Wong Yeong Yee, Vinod Ramachandran, Chai Suk Thin, See Pei Woei, Steven Lim Sung Wei, Shiidheshwar J Ravichandran, Walimuni DR Himashi Pemasiri, Anuradha Wanigaratne, Sisijaya Vinura Jayawardena, Michelle Chia Pei Ling, Fong Jing Hong, and Amery Ng Kai Xian. Their valuable contributions are gratefully acknowledged.
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]