A total of 363 COVID-19 patients responded to the survey. Of them, 124 (34.1%) claimed to have suffered from oral ulcers with a 95% CI (29–39%). The chi-squared test revealed no association between the prevalence of oral ulcers and age, sex, brushing habit, smoking, presence of chronic disease, or setting in which patients received COVID-19 treatment (Table 1). However, there was a significant association between corticosteroid use and the prevalence of oral ulcers (χ² = 4.22, p = 0.04). Corticosteroid use increased the risk of oral ulcer occurrence by 61% (OR = 1.61 with 95% CI (1.02–2.5), while other treatments had no effect (Fig. 1).
Table 1
Patient’s characteristics and their relationship to the incidence of mouth ulcer
Factor | Mouth ulcer | No mouth ulcer | Total | p-value |
n | % | n | % | n | % |
Age (in years) | < 30 | 48 | 13.2% | 96 | 26.4% | 144 | 39.7% | .96 |
30–59 | 70 | 19.3% | 132 | 26.4% | 202 | 55.6% |
60+ | 6 | 3% | 11 | 1.7% | 17 | 4.7% |
Sex | Male | 34 | 9.4% | 71 | 19.6% | 105 | 29% | .64 |
Female | 90 | 46.3% | 168 | 24.8% | 258 | 71% |
Symptoms | Asymptomatic | 0 | 0% | 5 | 1.4% | 5 | 1.4% | .17 |
Symptomatic | 124 | 34.2% | 234 | 64.5% | 358 | 98.6% |
Brushing | No | 13 | 3.6% | 40 | 11% | 53 | 14.6% | .11 |
Yes | 111 | 30.6% | 199 | 54.8% | 310 | 85.4% |
Smoking | Non-smoker | 107 | 29.5% | 206 | 56.7% | 313 | 86.2% | .98 |
Smoker | 17 | 4.7% | 33 | 9.1% | 50 | 13.8% |
Chronic diseases | No | 92 | 25.3% | 192 | 52.9% | 284 | 78.2% | .17 |
Yes | 32 | 8.8% | 47 | 12.9% | 79 | 21.8% |
Location of COVID-19 treatment | Home | 116 | 32% | 226 | 62.3% | 342 | 94.2% | .8 |
Hospital | 5 | 1.4% | 9 | 2.5% | 14 | 3.9% |
| ICU | 3 | .8 | 4 | 1.1% | 7 | 1.9% | |
P: chi-squared tests * P < 0.05 (significant) |
Concerning the association between COVID-19 symptoms and oral ulcers, the Chi-squared test revealed that only abdominal pain, shortness of breath, and diarrhea were associated with oral ulcer occurrence (p < 0.05) (Fig. 2).
The logistic regression model, using the equation, Log (OR oral ulcer) = -1.08 + 0.46 Shortness of breath (yes) + 1.17 Abdominal pain (yes) + 0.68 congestion (yes) showed that for each unit increase in the presence of symptoms (shortness of breath, congestion, and abdominal pain), the log (OR of oral ulcer) increased, holding all other covariates constant. This was applied using the odds ratio of having shortness of breath, which increased the probability of oral ulcer by a factor of 1.58 with 95% CI (1.03–2.48). The odds ratio of having abdominal pain elevated the probability of oral ulcer by a factor of 3.23 with 95% CI (1.29–8.58), while the odds ratio of having congestion during COVID-19 infection increased the probability of oral ulcer by a factor of 1.97 with 95% CI (1.26–3.12) (Table 2).
Table 2
The logistic regression model for incidence of mouth ulcer with coefficient, p-values, odds ratios, and 95% CI.
Variables | Odds ratio (95%CI) | Coefficient(B^) | Adjusted Odds ratio(95%CI) | p-value |
intercept | - | -1.08 | - | .02* |
Shortness of breath | 1.73 (1.12–2.69) | 0.46 | 1.58(1.03–2.48) | .04* |
Abdominal pain | 3.38 (1.35–8.78) | 1.17 | 3.23(1.29–8.58) | .014* |
Congestion | 2.13 (1.37–3.33) | 0.68 | 1.97(1.26–3.12) | .003* |
* P < 0.05 (significant) |
The multiple linear regression model demonstrated that the pain intensity score increased with age and the presence of abdominal pain but decreased in line with brushing habits, with admission to the hospital and/or the intensive care unit (ICU). Overall, the model could interpret 17.2% of the oral ulcer pain scale (R2 = 0.17, R2-adjusted = 0.14, F = 4.7, and p < 0.001) (Table 3).
Table 3
The logistic regression model for presence of ulcer with coefficient, p-values, odds ratios, and 95% CI.
Model | Estimate(B) with 95%CI | SE | t-test | p-value |
Intercept | 1.49(-0.23, 3.21) | .87 | 1.721 | .09 |
Age | 0.06 (0.02, 0.11) | .02 | .84 | .005* |
Brushing | -2.1 (-3.78, -0.42) | .85 | .47 | .015* |
Isolation (Hospital) | -1.27(-4.15, 1.62) | .46 | 0.87 | .39 |
Isolation (ICU) | -4.4 (-7.74, -1.07) | .68 | 2.61 | .01* |
Abdominal pain | 2.08 (0.31, 3.85) | .89 | .34 | .02* |
* P < 0.05 (significant) |
By applying the following equation, where oral ulcer pain scale = 1.49 + 0.06age-2.1 brushing habits (yes) − 1.27 isolation (hospital) -4.4 isolation (ICU) + 2.08 abdominal pain (yes), we could detect that the pain intensity score of the oral ulcer increased by 0.06 with each year of age with other variables remaining constant, 95% CI (0.02, 0.11), p = 0.005. In addition, the presence of abdominal pain increased the oral ulcer pain scale by 2.08 with 95% CI (0.31, 3.85), with other variables remaining constant, p = 0.02.
Good brushing habits decreased the oral ulcer pain scale by 2.1 with 95% CI (-3.78, -0.42), holding other variables constant (p = 0.015). Moreover, ICU admission decreased the oral ulcer pain scale score by 4.4 with 95% CI (-7.74, -1.07), p = 0.01, with other variables remaining constant. Hospital admission alone showed a non-significant decrease of 1.46 with 95% CI (-4.15, 1.62), p = 0.39.