A hospital-based comparative cross-sectional study was conducted at MKSH from January 11 to March 10, 2022. MKSH is located in Mettu Town, which is 600 km away from Addis Ababa, the capital city of Ethiopia, to the southwest. This hospital serves about 2.5 million people from different regions of the country. The hospital's health service covers the outpatient department, inpatient services, critical care, emergency intervention unit, and different clinics, such as antenatal care clinics, delivery services, tuberculosis clinics, ART clinics, and ophthalmology clinics. Currently, about 1533 HIV-infected adult clients are taking a different combination of ART drugs at this hospital.

Semi-structured and interviewer-administered questionnaires were used to collect data on socio-demographic and behavioral characteristics of the study participants. Data such as the duration of HIV disease since diagnosis, ART duration and regimen, viral load, CD4 count, and WHO clinical stage of HIV infection were obtained from the patient's medical record. Baseline clinical data were collected from medical records at the time of enrollment in ART programs.

Participant height was meticulously measured using a stadiometer, without shoes, to the nearest centimeter. To ensure accuracy, each individual stood upright against the stadiometer with their head, shoulders, buttocks, and heels firmly in contact with the scale, maintaining a level gaze straight ahead. Weight was precisely recorded using a calibrated Seca scale (Seca gmbh & co. kg, Germany) positioned on a level surface. The scale's accuracy was verified daily with known weights. Each participant stood barefoot, wearing light clothing, and their weight was documented to the nearest 0.1 kg. Finally, the Body Mass Index (BMI) was calculated for each participant using the established formula: weight in kilograms divided by height in meters squared (kg/m²).

Blood pressure (BP) was measured using a mercury sphygmomanometer and recorded to the nearest millimeter of mercury (mmHg). To ensure accurate readings, a standardized protocol was followed. Participants sat quietly for 5 min before the first measurement and again for 3 min before repeating it. The participant's arm rested comfortably with their palm facing up, supported at heart level, and legs uncrossed. They were encouraged to sit relaxed and comfortably. Two readings were taken from the right arm with a 3 min interval. If the difference between the two readings was less than 5 mmHg, the average was recorded as the final BP. For discrepancies exceeding 5 mmHg, a third measurement was taken, and the average of all three readings became the final BP.

Trained healthcare workers conducted HIV testing on participants in the comparison group using a blood-based rapid diagnostic kit, following the national algorithm for HIV testing. This approach aligns with the World Health Organization's guidelines for provider-initiated HIV testing and counseling (PITC). Before testing, each participant provided verbal consent and received counseling adapted to the WHO's PITC guidelines (29). Only participants testing negative for HIV were included in the comparison group.

A resting 12-lead body surface electrocardiography was performed using EDAN SE-1200Express electrocardiograph machine for each study participant. The ECG record was taken by a radiologist at a standardization of 1 mv = 10 mm and a paper speed of 25 mm/sec. Metallic ornaments and clothing above the waist were removed while respecting participant privacy. The skin was prepared for the electrode placement by shaving hair, rubbing with alcohol swabs, and the area dried with gauze pads. The cardiac gel was applied to electrode placement areas, and the electrodes were applied to the ankle of the lower legs, wrist of the forearms, and pericardial area. ECG was recorded in a quiet room with instructions to breathe normally, relax, remain calm and close eyes. The printed ECG record was reviewed and interpreted by a cardiologist and classified according to the Minnesota Code classification system (30).

Following data cleaning and coding, the information was entered into Epi Data (version 4.6.0.4) and exported to SPSS (version 25.0) for analysis. We explored the exported data to identify outliers and missing values. Categorical data were summarized as frequencies and percentages, with chi-square tests used to identify statistically significant differences between HIV-infected and HIV-negative groups. Continuous data were presented as means and standard deviations, with independent-samples t-tests employed to assess significant mean differences between the two groups. Bi-variable and multivariable logistic regression analyses were conducted to evaluate the influence of various factors on electrocardiographic findings. Variables exhibiting p-values < 0.25 in bi-variable analyses were included in the multivariable model. Backward selection was used to refine the multivariable model, and variables retaining p-values < 0.05 were deemed significantly associated with ECG abnormalities. The strength of association between independent and dependent variables was expressed using odds ratios with 95% confidence intervals.

The initial English questionnaires were translated into Afan Oromo and Amharic by qualified professionals and then back-translated to ensure consistency with the original meaning. To optimize clarity and accuracy, a pre-test with 5% of the planned sample size was conducted at Bedelle Hospital. Feedback from this pre-test guided further questionnaire refinements.

Data collectors and supervisors received comprehensive training to ensure standardized data collection and enhance reproducibility. All equipment was calibrated before each measurement to guarantee precision. An independent cardiologist, blinded to participants’ HIV status, interpreted all electrocardiograms, minimizing potential bias.

A total of 202 study participants including 101 HIV-infected patients and age-sex-matched 101 HIV-negative individuals for comparison were interviewed and underwent electrocardiography. The ECG record of 10 study participants were excluded due to poor quality. The analysis included data from 96 participants in each group.

Sixty-seven (69.8%) of the study participants were females in both HIV-infected and HIV-negative group (P = 1). The mean age was 38.9 years (SD ± 9.4) for HIV-infected group and 38.3 years (SD ± 10.17) for HIV-negative individuals (P = 0.681), with a minimum age of 20 and a maximum of 64 years. A higher proportion of HIV-infected individuals lived in urban areas (77.1%) compared to the HIV-negative group (61.5%, P = 0.019). Forty-four (45.8%) of HIV-infected and 73 (76%) of HIV-negative participants were married (Table 1).

Among the participants, 15.6% of HIV-infected and 10.4% of HIV-negative participants were smokers (P = 0.284). Similarly, there was no significant difference in the proportion of participants who reported alcohol drinking (13.5% in HIV-infected vs. 18.8% in HIV-negative; P = 0.327) or khat chewing (5.2% in HIV-infected vs. 10.4% in HIV-negative; P = 0.179) between the two groups. Physical activity levels were also comparable, with 53.1% of HIV-infected and 46.9% of HIV-negative participants categorized as physically active (P = 0.386).

The mean weight of HIV-infected and HIV-negative respondents were 60.6 ± 10.39 and 62.42 ± 6.5 kg respectively (P = 0.158), while their respective mean height was 1.66 ± 0.05 and 1.67 ± 0.06 meters (P = 0.087). Similarly, the mean BMI of HIV-infected and HIV-negative respondents were 21.99 ± 3.4 and 22.3 ± 2.04 Kg/m² respectively (P = 0.435). The mean of diastolic blood pressure and systolic blood pressure in HIV-infected and HIV-negative participants were (73.59 ± 6.87 vs. 74.58 ± 7.52 mmHg; P = 0.343) and (112.92 ± 9.33 vs. 115.21 ± 9.51 mmHg; P = 0.094) respectively.

HIV-infected participants reported an average infection duration of 10.43 years (SD ± 4.22) and had been on ART for an average age of 8.87 years (SD ± 3.81). The mean recent CD4 cell count and viral load were 573.5 ± 278 cells/mm³ and 795.4 ± 2,216.4 viral copies/ml respectively, while the recent viral load was undetectable for 61% of HIV-infected participants. All HIV-infected participants were on ART medication, with 16.7% receiving regimens containing PIs (Table 2).

In this study, 47 (49%) of the HIV-infected participants exhibited at least one ECG abnormality, compared to only 19 (19.8%) of the HIV-negative participants. This difference in prevalence was found to be statistically significant, indicating that ECG abnormalities are more common in HIV-infected individuals (χ² = 18.1, P < 0.001) (Figure 1).

When comparing the types of ECG abnormalities, the prevalence of arrhythmia was similar between HIV-infected and HIV-negative participants. However, the mean heart rate was significantly higher in HIV-infected individuals. ST-segment abnormalities, T wave abnormalities, and increased QTc interval were significantly higher in the HIV-infected group compared to the HIV-negative group. Conduction abnormalities, ventricular hypertrophy and QRS axis deviations were not significantly different between the two groups. Other ECG abnormalities such as pathologic Q wave, low voltage QRS complex, poor R wave progression, and tall T wave were exclusively seen in HIV-infected individuals (Table 3).

Binary logistic regression was employed to determine factors associated with the ECG abnormality in HIV-infected participants. The final model fitness was checked by using Hosmer-Lemeshow goodness-of-fit test.

In bivariable logistic regression; residence, smoking status, physical activity status, HIV infection duration since diagnosis, ART regimen, recent viral load, recent CD4 level and WHO clinical stage of participants were associated with ECG abnormalities at a P-value of <0.25. Multivariable logistic regression using backward LR was done to identify the best predictors of ECG abnormalities. The collinearity of independent variables was checked by running pseudo linear regression, and fitness of the final model was evaluated by using Hosmer-Lemeshow goodness-of-fit test. Accordingly, the finding from multivariable logistic regression showed that smoking status, ART regimen and recent CD4 level were significantly associated with ECG abnormalities at a significance level of <0.05.

Compared to non-smokers, respondents who were smokers were 3.7 times more likely to have ECG abnormality (AOR = 3.7, 95%CI: 1.03–13.6). HIV infected respondents who were on PIs containing ART regimen had a 3.6 times higher chance of developing abnormal ECG findings than those who were on non-PIs containing ART regimen (AOR = 3.6, 95%CI: 1.02–13.1). HIV patients with recent CD4 cell count <350 cells/mm³ were 3.2 times more likely to have ECG abnormality than those with CD4 cell count ≥350 cells/mm³ (AOR = 3.2, 95%CI: 1.22–8.49) (Table 4).

After controlling for sex, smoking status, alcohol consumption, and physical activity, HIV infection was found to be independently linked to ECG abnormalities (AOR = 3.8; 95%CI: 1.99–7.38). HIV-infected individuals were 3.8 times more likely to experience ECG abnormalities compared to those without HIV infection (Table 5).

Since it is cross sectional study, it is difficult to determine the precise relationship between variables and ECG abnormalities. In addition, biochemical tests such as serum electrolytes, lipid profiles and cardiac inflammatory markers were not done due to financial constraints. We were also unable to determine the actual time of HIV infection so the exact duration of HIV infection was not known.