Abstract

Introduction

Many researchers utilize artificial intelligence (AI) to aid their research endeavors. This study seeks to assess and contrast the performance of three sophisticated AI systems, namely, ChatGPT, Gemini, and Perplexity when applied to an examination focused on knowledge regarding research publication.

Methods

Three AI systems (ChatGPT-3.5, Gemini, and perplexity) were evaluated using an examination of fifty multiple-choice questions covering various aspects of research, including research terminology, literature review, study design, research writing, and publication-related topics. The questions were written by a researcher with an h-index of 22, and it was later tested on two other researchers with h-indices of 9 and 10 in a double-blinded manner and revised extensively to ensure the quality of the questions before testing them on the three mentioned AI systems.

Results

In the examination, ChatGPT scored 38 (76%) correct answers, while Gemini and Perplexity each scored 36 (72%). Notably, all AI systems frequently chose correct options significantly: ChatGPT chose option (C) correctly 88.9% of the time, Gemini accurately selected option (D) 78.9% of the time, and Perplexity correctly picked option (C) 88.9% of the time. In contrast, other AI tools showed minor agreement, lacking statistical significance, while ChatGPT exhibited significant concordance (81-83%) with researchers' performance.

Conclusion

ChatGPT, Gemini, and Perplexity perform adequately overall in research-related questions, but depending on the AI in use, improvement is needed in certain research categories. The involvement of an expert in the research publication process remains a fundamental cornerstone to ensure the quality of the work.

Introduction

The work of John McCarthy is the foundation of modern artificial intelligence (AI) research. In 1956, at Dartmouth College, he introduced the phrase "artificial intelligence," marking the inception of formal AI research [1]. The emergence of AI was an innovative technological frontier, promising transformative impacts across diverse sectors. Recent years have witnessed significant strides in the AI domain, particularly in the refinement of chatbot technology. An increasingly prevalent notion suggests that AI, having surpassed human capabilities in several domains, holds promise for substantial advancements in the realm of research publications. AI stands poised to augment research writing, the accuracy of information retrieved, and referencing, thereby potentially revolutionizing the field [2].

Over the past few years, a multitude of AI tools have become readily accessible, providing a diverse array of services and functionalities. A notable instance of such an AI system is ChatGPT, an advanced language model crafted by OpenAI. It underwent training using a vast array of textual materials gathered from websites, literature, and diverse sources, engaging in language modeling tasks to enhance its capabilities. This attribute sets it apart as one of the most expansive and resilient language models ever devised, integrating an astonishing 175 billion parameters [3,4]. An additional AI system that has attracted attention is Gemini, previously identified as Google Bard, which is an AI-driven information retrieval apparatus with a sophisticated chatbot that utilizes a "native multimodal" approach to effectively process and adjust to various types of data like video, audio, and text [5,6]. Perplexity AI stands as an AI-powered research and conversational search engine, adept at responding to queries through the utilization of natural language predictive text. It synthesizes answers from web sources, accompanied by citations through embedded links within the text response [7]. Many researchers are known to utilize chatbots as aids in their research endeavors.

This study seeks to assess and contrast the performance of sophisticated AI systems—namely, ChatGPT, Gemini, and Perplexity—when applied to an examination focused on knowledge regarding research publication. It also aims to shed light on the current state of AI integration within the research publication process and identify opportunities for further development.

Methods

In this comparative investigation, we evaluated the performance of three distinct AI systems: ChatGPT-3.5, Gemini, and Perplexity. The assessment comprised 50 multiple-choice questions, each offering four options (A-D). The questions spanned various domains including eleven research terminology queries, six literature review inquiries, twelve study design probes, twelve research writing assessments, and nine publication-related investigations.

Initially, a researcher with an h-index of 22, identified as the second author in the manuscript, composed a set of sixty multiple-choice questions. Subsequently, two other researchers with h-indices of 14 and 16, mentioned as authors seven and ten respectively, underwent the examination in a double-blinded fashion. Following this phase, all three researchers collaborated to review and analyze both questions and answers. Ten questions were excluded due to their lack of clarity, leaving a total of fifty questions selected for the final examination version. These selected questions were unanimously agreed upon by the researchers as informative indicators of knowledge within the realm of research and its associated intricacies.

The questions were then uniformly inputted into each of the AI systems in March 2024, following a standardized protocol. This protocol involved initiating interactions with the AI systems by introducing a prompt starting with "Hello." Subsequently, each AI system received the same directive: "Please select the correct answer for the following multiple-choice questions." The questions were directly transcribed from a prepared Word document, and the AI-generated responses were recorded in an Excel spreadsheet. Statistical analysis was performed using Statistical Package for the Social Sciences (SPSS) version 27.0, with a significance level set at p < 0.05. Chi-square (Fisher's Exact Test) was employed for data analysis.

During the literature review phase of the present study, papers were selectively included from reputable journals and omitted those published in predatory journals, adhering to the criteria delineated in Kscien’s list [8].

Results

In the examination, ChatGPT demonstrated slightly higher accuracy with a total of 38 correct answers (76%), compared to 36 correct answers (72%) by both Gemini and Perplexity. Notably, Researcher 2 excelled in terminology and literature review questions, with 15 correct answers (88.23%), surpassing ChatGPT and Gemini, with 13 correct answers (76.47%). In research writing, Perplexity, along with Researcher 1 and Researcher 2, led with 10 correct responses (83.3%). Additionally, Researcher 1 exhibited the highest accuracy in research publication, with 9 correct responses (100%), outperforming ChatGPT and Researcher 2, who achieved 7 correct responses (77.78%) (Supplementary 1).

In the examination comparing AI tools and two researchers' accuracy in identifying correct answers, researchers demonstrated superior accuracy compared to AI tools. For example, in questions where the correct answer was C, Researcher 2 achieved a perfect 100% accuracy, outperforming ChatGPT, Perplexity, and Gemini, which scored 88.9%, and 77.8% respectively. Notably, all AI systems significantly chosen the correct options. For instance, ChatGPT correctly identified option C 88.9% of the time, Gemini correctly chose option D 78.9% of the time, and Perplexity accurately selected option C 88.9% of the time (Table 1).

In comparing AI tools and researchers' performance, significant agreement was noted with ChatGPT. For instance, out of 43 questions where researcher 1 agreed on the correct answer, ChatGPT agreed in 35 cases (81.4%) and disagreed in only 8 answers (18.6%). However, the comparison with the other two AI tools showed no significance but a slight alignment with the researchers' agreement on the correct answers (Table 2).

Discussion

The imitation of human intelligence functions by machines, most commonly computer systems, is referred to as AI. It involves acquiring knowledge (gaining information and understanding rules for its utilization), logical deduction (applying rules to arrive at rough or precise outcomes), and self-adjustment. In addition, AI endeavors to develop systems capable of executing tasks traditionally associated with human intelligence, including decision-making, speech recognition, language translation, and visual perception, among various others [9]. Although AI language models have been in development for years, the general population's understanding of AI's potential and use has increased dramatically recently. The academic community has already embraced language-based AI, and numerous researchers utilize chatbots as aids in their research. These bots assist in structuring ideas, offering feedback on their work, and aiding in referencing and summarizing the existing research literature [2,10,11].

Kacena et al. demonstrated that the utilization of AI, particularly ChatGPT, reduced the time invested in crafting review articles. However, it yielded the highest similarity indices, indicating a greater probability of plagiarism. In addition, they reported that ChatGPT possesses the ability to swiftly scour the internet and evaluate potential sources, potentially accelerating the literature review process. In the current study, the performance of ChatGPT regarding the principle of literature review questions showed a high performance, and Gemini scored just as high, further supporting the finding of the previous study [12].

Salvagno et al. reported that AI may soon be leveraged for the automated production of figures, tables, and supplementary visual components within manuscripts. This utilization could facilitate data summarization and contribute to manuscript lucidity [13]. However, the current study demonstrated that the AI systems had different scores, and their performance was influenced by the different categories they were tested on, which means that identifying the strengths and weaknesses of the currently available AIs is paramount in choosing which AI system will aid in research publications rather than hindering and jeopardizing the integrity of the research paper, For instance, Kacena et al. showcased that 70% of the references were incorrect when an AI only method was applied to writing research papers, raising controversy if these AI tools should even be used as aid in that regard [12]. The present study showed that Gemini performed poorly by only getting half of the questions wrong in the research writing principles questions. In addition, Perplexity was shown to perform poorly on principles of publication-related questions, and ChatGPT exhibited subpar performance in research terminology inquiries, further supporting the notion that leveraging AI use is dependent on recognizing their limitations in the field of research.

Concerns about biases in AI systems, stemming from their training data, are widely recognized as a significant challenge. Research indicates that AI models can perpetuate biases and exhibit skewed behavior, replicating existing discriminatory patterns. Addressing these biases is crucial and requires the implementation of effective strategies prioritizing fairness and justice during development. This is particularly important in research, where ensuring impartiality is paramount. Responsible use of advanced language models like ChatGPT, Gemini, and Perplexity is essential, given the ethical dilemmas they pose, including the potential for misinformation and emotionally persuasive content. Proactive steps are needed to mitigate these risks and promote responsible usage. Additionally, the use of AI in content generation raises concerns about unintentional plagiarism, as systems may reproduce text without proper citation. While AI tools may increase publication output, there may not be a corresponding increase in expertise or experience among researchers [3,12].

Several studies have investigated the comparison of AI and human capabilities across various domains. Long et al. noted a remarkable level of accuracy in AI, ranging from 90% to 100% when evaluating its performance against specialized doctors' diagnostic and treatment decisions for congenital cataracts [14]. Additionally, Rajpurkar et al. discovered consistency in results between AI and radiologists, particularly in diagnosing chest radiographs [15]. However, there is limited available data on the comparison of AI and human performance in research principles. In this study, the comparison between AI tools and human performance regarding predetermined correct answers on research principles revealed a significant agreement (80-85%) between ChatGPT and researchers.

One of the limitations of our study is that we evaluated only three AI systems in comparison to the vast and increasing number of AI tools becoming available in these times. In addition, a larger number of questions will lead to a more comprehensive understanding of the strengths and weaknesses of these AI systems in the field of research and their utilities in that regard.

Conclusion

ChatGPT, Gemini, and Perplexity perform adequately overall in research-related questions, but depending on the AI in use, improvement is needed in certain research categories. The involvement of an expert in the research publication process remains a fundamental cornerstone to ensure the quality of the work.

Declarations

Conflicts of interest: The author(s) have no conflicts of interest to disclose.

Ethical approval: Not applicable.

Patient consent (participation and publication): Not applicable.

Funding: The present study received no financial support.

Acknowledgements: None to be declared.

Authors' contributions: RQS and SHM were major contributors to the conception of the study and the literature search for related studies.  AMS, JOA, DSH, and AMS were involved in the literature review, the study's design, and the critical revision of the manuscript, and they participated in data collection. HAH, and YMM were involved in the literature review, study design, and manuscript writing. BAA, DSH, and RQS Literature review, final approval of the manuscript, and processing of the tables. RQS and SHM confirm the authenticity of all the raw data. All authors approved the final version of the manuscript.

Use of AI: AI was not used in the drafting of the manuscript, the production of graphical elements, or the collection and analysis of data.

Data availability statement: Note applicable.

Acknowledgement: Not applicable.

Abstract

Introduction

Due to indeterminate cytology, Bethesda III is the most controversial category within the Bethesda System for Reporting Thyroid Cytopathology. This study examines the characteristics and malignancy rates of thyroid nodules (TNs) classified as Bethesda III.

Methods

Data were collected by reviewing electronic medical records, encompassing demographic details, medical history, chief complaint, laboratory tests (including thyroid function tests), preoperative imaging, cytology results, management, and histopathology diagnosis.

Results

The majority of the cases were female (84.7%). Patients’ ages ranged from 15 to 71 years, with a mean of 42.9 ± 10.5 years. Regarding goiter grading, 37 cases (21.8%) were classified as G0, 62 (36.5%) as G1, 55 (32.3%) as G2, and seven (4.1%) as G3. Thyroid Imaging Reporting and Data Systems scoring categorized the nodules as TI-RADS 2 (5.3%), TI-RADS 3 (40%), TI-RADS 4 (38.2%) and TI-RADS 5 (9.4%).  The size of TNs on ultrasound ranged from 0.3 cm to 7.8 cm, with a mean size of 2.06 ± 1.3 cm. Adenoma was the most common diagnosis (40%), followed by thyroiditis (16.5%), papillary thyroid carcinoma (15.9%), and papillary thyroid microcarcinoma (15.9%). The nodules were predominantly benign (64.7%), while 35.3% were malignant. Patients with malignant nodules were younger than those with benign nodules (p=0.044). Benign nodules were significantly larger than malignant ones (p-value = 0.003).

Conclusion

One of three TNs with indeterminate cytology may be malignant. Patients with malignant nodules tend to be younger than those with benign nodules, and benign nodules are likely larger than malignant ones.

Introduction

The global incidence of thyroid nodules (TNs) is estimated to range between 20% and 60%, varying by gender, age, and geographic location [1]. Almost 90–95% of these nodules are benign and asymptomatic at diagnosis and remain so during follow-up [2]. However, the incidence of thyroid cancer, including papillary thyroid carcinoma (PTC) and papillary thyroid microcarcinoma (PTMC), has risen concurrently with the advancements in diagnostic technology and enhanced surveillance. Additionally, incidence-based mortality from thyroid cancer has increased, with an annual percent change of 1% [1,3]. With new ultrasound (U/S) technology and the widespread use of high-resolution scanners, detecting TNs has become much easier. However, for many sonographers, the primary challenge lies in accurately distinguishing malignant TNs from benign ones. To address this, certain U/S characteristics, such as unclear borders, micro-calcifications, irregular shapes, solid components, and internal echoes, are commonly used to assess the malignancy risk of nodules. Nonetheless, relying on any single characteristic alone is insufficient to accurately differentiate between malignant and benign nodules [4]. Fine needle aspiration cytology (FNAC) has become the standard modality for assessing thyroid nodular pathology [1]. In 2008, the Bethesda System for Reporting Thyroid Cytopathology (BSRTC) was introduced to standardize the cytological evaluation of TNs. The BSRTC categorizes diagnoses into six classes with progressively higher suspicion for malignancy: nondiagnostic (Class I), benign (Class II), atypia of undetermined significance /follicular lesion of undetermined significance (AUS/FLUS) (Class III), follicular neoplasm/oncocytic cell neoplasm (Class IV), suspicious for malignancy (Class V), and malignant (Class VI) [5]. The most controversial category within the BSRTC is AUS/FLUS due to indeterminate cytology. Despite the routine use of cytological examination in evaluating TNs, which has reduced the overall number of patients needing diagnostic surgery, a significant percentage still undergo surgery to obtain a definitive histological diagnosis [1]. The reported malignancy rates for AUS and FLUS exhibit considerable variability [3]. This study examines the characteristics and malignancy rates of TNs classified as Bethesda III. The referenced studies have been checked to avoid citing non-peer-reviewed data [6].

Methods

Study design

This retrospective, single-center, cross-sectional study was conducted at Smart Health Tower between August 2024 and September 2024. The patients gave verbal informed consent to publish their data in this study. The ethical board at Kscien organization approved the study with approval number 26 on August 2024.

Data collection

Data were collected by reviewing electronic medical records, encompassing demographic details, medical history, chief complaints, laboratory tests (including thyroid function tests), preoperative imaging, FNAC results, management, and final histological diagnosis.

Eligibility criteria

The study included patients with TNs classified as Bethesda III on FNAC who subsequently underwent surgery for a definitive histopathological diagnosis. Patients with incomplete medical documentation, including clinical, radiological, and FNAC data, as well as those with recurrent or a history of thyroid cancer, were excluded.

Statistical Analysis

The data were collated in a Microsoft Excel (2021) sheet and then transferred into version 27 of Statistical Package for Social Sciences (SPSS). The chi-squared and Fisher's exact tests were used to analyze categorical data with independent samples t-test for quantitative variables. The data were presented as frequency, percentage, mean with standard deviation, and median with ranges. The level of significance was set at p-value <0.05.

Results

Patient demography and presentation

The study included 170 patients, of whom the majority were female (84.7%). Their ages ranged from 15 to 71 years, with a mean of 42.9 ± 10.5. In total, 157 patients (92.3%) were married, 11 (6.5%) were unmarried, and two individuals (1.2%) were either divorced or widowed. Most of the cases (65.9%) were housewives. Seven cases were smokers (4.1%), while 20 were passive smokers (11.8%), and one (0.6%) was an ex-smoker. The reason for the presentation was having a thyroid disease and visiting the hospital for follow-up in most of the cases (38.2%), followed by neck swelling (34.7%) and fatigue (10.0%) (Table 1). Regarding goiter grading, 37 cases (21.8%) were classified as G0, 62 (36.5%) as G1, 55 (32.3%) as G2, and seven (4.1%) as G3. Nine patients (5.3%) had no goiter grading available (Table 2).

Diagnosis and management

Thyroid function assessment revealed that the majority of the participants were euthyroid (51.2%). However, hypothyroidism and hyperthyroidism were observed in 24.7% and 17.1% of patients, respectively. Ultrasonography of the thyroid showed that 113 patients (66.5%) had solid nodules, while 32 (18.8%) presented with mixed (solid and cystic) nodules. Only one case (0.6%) had a purely cystic nodule. The content of the tumor in the remaining cases was unknown (14.1%). Thyroid Imaging Reporting and Data Systems (TI-RADS) scoring categorized the nodules as TI-RADS 2 (5.3%), TI-RADS 3 (40%), TI-RADS 4 (38.2%) and TI-RADS 5 (9.4%). It was unknown in 12 cases (7.1%). The size of the TNs on U/S ranged from 0.3 cm to 7.8 cm, with a mean of 2.06 ± 1.3 cm. The majority of the cases were managed by total thyroidectomy (67.6%), followed by lobectomy (17.6%), nodulectomy (12.4%), and isthmusectomy (2.4%). Adenoma (Figure 1) was the most common diagnosis, identified in 68 cases (40%), followed by thyroiditis (16.5%), PTC (15.9%) (Figure 2), and PTMC (15.9%) (Figure 3). Benign multinodular goiter (MNG) was diagnosed in seven individuals (4.1%), noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) in five patients (2.9%) (Figure 4), and follicular thyroid carcinoma (Figure 5) in four cases (2.3%). Other less common diagnoses included Graves’ disease in two patients (1.2%), a collision tumor consisting of PTMC and follicular adenoma in one individual (0.6%), and medullary thyroid carcinoma in another patient (0.6%). The nodules were predominantly benign (64.7%), while 35.3% were malignant (Table 2). 

Comparison of patient and tumor characteristics between benign and malignant TNs

In comparing patients' demography and clinical, and radiological characteristics between benign and malignant nodules, no significant differences were observed between the two groups concerning age group, gender, goiter grading, MNG (data not shown), thyroid function, or tumor features on U/S (p-value > 0.05). However, in general, patients with malignant nodules tended to be younger than those with benign nodules (p=0.044). TI-RADS classification differed significantly between benign and malignant nodules. Malignant nodules were significantly associated with TI-RADS categories 4 and 5, whereas benign nodules were predominantly associated with TI-RADS categories 2 and 3 (p-value < 0.001). There was no statistically significant difference in tumor size between benign and malignant nodules when categorized into groups (p-value = 0.053) (Table 3). However, when tumor size was not grouped, benign nodules were significantly larger than malignant ones (p-value = 0.003) (Table 4).

Discussion

Although FNAC is widely utilized, its diagnostic accuracy ranges from 80% to 99%. The ambiguity associated with indeterminate cytological outcomes often leads to uncertainty, and repeating FNAC following a non-diagnostic result remains controversial [7]. Thyroid pathology, especially TNs, primarily affects women. The majority of thyroid cancer patients are women, with an average age of 50 ± 15 years. However, men are more likely to develop aggressive forms of the disease, which are linked to a poorer prognosis [8]. The global incidence of thyroid cancer is on the rise. In 2020, the estimated incidence rate was 10.1 per 100,000 people for women and 3.1 for men, up from 6.1 and 1.9, respectively, in 2012 [8]. Regarding the correlation of gender and age with the risk of malignancy in TNs with indeterminate cytology, controversial findings have been reported [1,3,8,9]. Bessey et al. reported female gender as a risk factor for malignancy [9]; however, Cozzani et al. found that female gender was associated with well-differentiated thyroid cancer predominantly in younger individuals, but this difference diminished in patients over the age of 55 [1]. In contrast, Rano et al. found no effect of gender, race, ethnicity, and underlying thyroid disease on thyroid cancer [8]. Several studies have identified a minimal inverse correlation between patient age and malignancy, suggesting that younger patients with indeterminate nodules face a slightly higher risk of being diagnosed with thyroid cancer [1,8-10]. On the contrary, Dimitriadis et al. reported a similar mean age of 50 between both groups of patients with benign and malignant nodules [3]. In line with existing literature, most of the cases in the present study were female (84.7%); however, gender did not differ significantly between those with malignant and benign nodules. Additionally, no significant difference between malignant and benign nodules was observed when age was categorized into groups. Nonetheless, patients with malignant nodules tended to be younger, with a mean age of 40.76 years, compared to those with benign nodules, who had a mean age of 44.16 years. This finding supports the assumption that TNs in younger patients are more likely to be malignant.

Some scholars have identified iodine deficiency as an indirect risk factor for thyroid cancer. Iodine plays a crucial role in the synthesis of thyroid hormones, and its deficiency can lead to an increase in thyroid volume (goiter) and elevated thyroid-stimulating hormone production. Rano et al. reported a high incidence of goiter among their cases (63%), with malignant histology more commonly associated with MNG (71%) than a single nodule. They found no difference between benign and malignant tumors regarding nodule content (solid, cystic, mixed) on U/S [8]. In the present study, only seven cases (4.1%) had MNG, and contrary to the previous study, there was no association between MNG and an increased risk of malignancy. However, consistent with the findings of Rano et al. [8], nodule content did not differ based on tumor nature, whether benign or malignant.

The usefulness of nodule size as an independent predictor of malignancy remains controversial. Both the British Thyroid Association and the American Thyroid Association recommend total thyroidectomy for indeterminate lesions measuring ≥40 mm due to the associated increased risk of malignancy [11,12]. In a study by Dimitriadis et al., the average nodule size was comparable between benign and malignant subgroups, measuring under 4 cm (3.5 cm vs. 3 cm, respectively). The study also found that the likelihood of a nodule being malignant was similar regardless of its size, whether it was <4 cm or ≥4 cm (27% and 27.7%, respectively) [3]. In line with these findings, some other studies also identified no correlation between nodule size and the risk of malignancy [1,13,14]. Despite that, some others reported different findings. A study from Oxford found that approximately 37% of TNs classified as Bethesda III and measuring over 4 cm were malignant, which was significantly higher than the malignancy rate observed in nodules smaller than 4 cm [15]. Conversely, Cavallo et al. found that larger nodules had a lower malignancy rate and suggested that nodule size should not be considered an independent risk factor for malignancy [16]. 

Cozzani et al. reported that only 11.6% of the nodules in their study were larger than 4 cm. They attributed this finding to the extensive use of U/S examinations, including those conducted for screening purposes, in their area, which has a very high incidence of thyroid nodular pathology [1]. In the current study, only 8.8% of the cases had nodules greater than 4 cm, which may be due to the reason mentioned by Cozzani et al [1]. Our findings were consistent with the presumption that tumor size in benign nodules may be greater than that of malignant ones. 

The literature demonstrates considerable variability in the reported malignancy rates for AUS/FLUS TNs. In a study by Dimitriadis et al., the rate was about 34%, and in the 2017 National British Association of Endocrine and Thyroid Surgeons Audit report, it was 25.7% [3,17]. The Oxford group found that approximately one in four patients with AUS/FLUS cytology was diagnosed with thyroid cancer [15]. Similar results were observed in a systematic review that included 13 studies, revealing a malignancy rate of 22% for cases with Bethesda III cytology [18]. Additionally, malignancy rates of 34.9% and 39% have also been reported [8,19]. In the current study, the malignancy rate among Bethesda III nodules was 35.3%, comparable to what has been reported in the literature.

In a study by Bresler et al., 9% of Bethesda III nodules were histologically malignant, with 50% of these being PTC and 30% PTMC [20]. Another study found PTC as the predominant histological type among malignant TNs [8], while Finlayson et al. reported follicular carcinoma as the predominant type [21]. In this study, PTC and PTMC were the most common cancer types among the malignant nodules, with equal incidence (15.9%).

The primary objective of thyroid surgery for nodules classified as AUS/FLUS is to achieve a definitive histological diagnosis while ensuring complete removal of the pathological nodule. This approach aims to facilitate optimal surgical and medical management, thereby minimizing the risks associated with excessive surgical intervention and related adverse events [1]. Surgical options for TNs, such as total thyroidectomy, lobectomy, or nodulectomy, are influenced by several factors. These include risk factors indicating a higher likelihood of malignancy (such as nodules larger than 4 cm, a family history of neoplasia, and a history of radiation exposure), U/S characteristics, cytological category, and molecular testing. Additionally, these risk factors should be considered in conjunction with the patient’s preferences, the presence of contralateral nodularity, possible coexisting hyperthyroidism, and any comorbid conditions [22]. It is important to note that total thyroidectomy is no longer universally recommended for all differentiated carcinomas larger than 1 cm. According to the 2015 ATA Guidelines, lobectomy may be an adequate initial treatment for differentiated carcinomas smaller than 4 cm, NIFTP of any size, minimally invasive follicular carcinomas, and encapsulated or intrathyroidal variants of papillary carcinomas [1,22]. Another study recommended that nodulectomy may be an appropriate option for managing large, solitary TNs and small suspicious nodules or microcarcinomas [23]. In the present study, 67.6% of the cases underwent total thyroidectomy, with lobectomy in 17.6%, nodulectomy in 12.4%, and isthmusectomy in 2.4%. The present study's limitations included a small sample size and the retrospective nature of data collection, which may have led to the omission of important data, such as details on ultrasonography.

Conclusion

One of three TNs with indeterminate cytology may be malignant. Patients with malignant nodules tend to be younger than those with benign nodules, and benign nodules could be larger. While total thyroidectomy is common, lobectomy and nodulectomy may be viable alternatives for specific cases, emphasizing the need for individualized treatment.

Declarations

Conflicts of interest: The author(s) have no conflicts of interest to disclose.

Ethical approval: The study's ethical approval was obtained from the scientific committee of the Kscien Organization for Scientific Research.

Patient consent (participation and publication): Verbal informed consent was obtained from patients for publication.

Source of Funding: Smart Health Tower.

Role of Funder: The funder remained independent, refraining from involvement in data collection, analysis, or result formulation, ensuring unbiased research free from external influence.

Acknowledgements: None to be declared.

Authors' contributions: AMS and AMA were significant contributors to the conception of the study and the literature search for related studies.  RMA, AJQ, ROM and RJR were involved in the literature review, the study's design, and the critical revision of the manuscript, and they participated in data collection. HOA, HOB, and AAQ were involved in the literature review, study design, and manuscript writing. HAA and SHH Literature review, final approval of the manuscript, and processing of the tables. RMA was the pathologist who performed the histopathological diagnosis. HOA and RMA confirm the authenticity of all the raw data. All authors approved the final version of the manuscript.

Use of AI: AI was not used in the drafting of the manuscript, the production of graphical elements, or the collection and analysis of data.

Data availability statement: Not applicable.

Abstract

Introduction

Hospitals are high-risk environments for infections. Despite the global recognition of these pathogens, few studies compare microorganisms from community-acquired and hospital-acquired infections (HAIs). This study compares these microorganisms and explores their relationship with patients' comorbidities and socio-demographic factors.

Methods

This retrospective cross-sectional study was conducted at Smart Health Tower, Iraq, from January to December 2023, focusing on patients with community-acquired infections and HAIs. Data were extracted from microbiology laboratory records, including blood cultures, urine samples, and other body fluids, with patients classified based on CDC and IDSA guidelines. Bacterial identification combined conventional methods and the BD Phoenix™ M50 system, while antibiotic susceptibility was tested using the Kirby-Bauer method and the same automated system. Statistical analysis of resistance patterns utilized SPSS version 25, with significance set at p ≤ 0.05.

Results

In this study of 2,157 participants, 1,303 (60.4%) were male, with microbial growth observed in 1,177 cases (54.6%). Notably, 41.1% of females and 52.1% of males showed no growth (p < 0.001). The mean age was 43.62 ± 23.3 years. Wound samples had the highest growth rate (91.2%), while body fluids had the lowest (33.0%) (p < 0.001).  The overall multidrug resistance rates were 62.5% for community-acquired infections and 79.3% for HAIs. Patients with pre-existing comorbidities demonstrated significantly higher rates of hospital-acquired infections (p < 0.05).

Conclusion

Multidrug-resistant isolates are more prevalent in HAIs than in community-acquired infections, highlighting the need for enhanced surveillance to optimize antibiotic use and control HAIs through early detection of resistance.

Introduction

Hospitals represent a potentially hazardous environment due to various virulent pathogens introduced by admitted patients from the community. These patients are subsequently exposed not only to the hospital's endemic flora but also to microorganisms carried by other ill individuals [1]. This occurs due to a compromised immune defense and colonization by resistant organisms [2]. Hospital-acquired infections (HAIs) are a frequent occurrence in healthcare facilities globally, with their prevalence exceptionally high in resource-limited developing countries [3]. The extensive use of broad-spectrum antibiotics in hospitals creates an intense selective pressure, fostering the emergence of antibiotic-resistant bacteria and complicating the treatment of these infections. As a result, HAIs have been recognized as a severe public health issue for over a century, contributing to poor health outcomes and significantly impacting the quality of healthcare delivery [4]. 

Hospital-acquired infections most commonly manifest as urinary tract infections, respiratory tract infections, circulatory system infections, and surgical site infections [5]. A World Health Organization report covering 55 hospitals across 14 countries found that 8.7% of hospitalized patients developed HAIs, with the highest prevalence observed in the Eastern Mediterranean Region and lower rates in the Western Pacific [5]. The prevalence of HAIs has been reported at approximately 5% in North America and parts of Europe while reaching up to 40% in some areas of Asia, Latin America, and Africa [6]. A European study reported the prevalence of HAIs to be approximately 2.9%. Several factors contribute to the occurrence of HAIs, including medical interventions, substandard hospital environments, and inadequate personal hygiene practices among both hospital staff and patients [7]. However, the primary driver of HAIs is the failure to adhere to health and safety protocols in healthcare settings. While it is impossible to eliminate HAIs, even in highly advanced hospitals, strict adherence to established standards and guidelines can significantly reduce or manage their occurrence, especially in regions such as Africa [6]. In modern healthcare, where technological advancements and high expectations for quality care prevail, it is critical to thoroughly examine the frequency and underlying causes of HAIs. The absence of accurate data on the prevalence of HAIs poses significant challenges to executing these control measures, leading to increased healthcare costs for both health systems and patients [8].

Despite the global recognition of these pathogens, limited studies have compared microorganisms from both community and hospital settings; therefore, the current study aims to fill this gap by comparing microorganisms isolated from community-acquired and HAIs. It also seeks to explore the relationship between these infections and patients' comorbidities and socio-demographic factors.

Methods

Study design and setting

This retrospective cross-sectional study was conducted at Smart Health Tower, Iraq, between January 2023 and December 2023. It included patients from various departments of the hospital, with infections categorized as either community-acquired or HAIs. The Kscien Organization approved the study for Ethical Approval, reference number 24/No. 27, ensuring all ethical guidelines were followed throughout the study.

Sample collection and study population

Data were meticulously extracted from the records of patients who had their samples processed in the microbiology laboratory. Inclusion criteria encompassed all available clinical samples, including blood cultures, urine samples, sputum and bronchoalveolar lavage, wound swabs, and other body fluids. Patients were classified into either the CAI or HAI group based on guidelines from the Centers for Disease Control and Prevention and the Infectious Diseases Society of America. The CAIs were defined as infections present at the time of hospital admission or within 48 hours of admission, with no history of recent healthcare exposure, such as hospitalization within the previous 90 days. In contrast, HAIs were defined as infections that developed 48 hours or more after hospital admission and were associated with invasive procedures or prior healthcare exposure [9]. Patients with incomplete data were excluded to ensure the accuracy and reliability of the study findings.

Bacterial identification

Bacterial identification was conducted using conventional methods and the BD Phoenix™ M50 automated identification and susceptibility testing system, specifically tailored to the diverse range of clinical samples processed during the study. Blood cultures were incubated in the BD BACTEC™ automated blood culture system, following established protocols, for up to five days to detect the growth of bacteria or fungi, with positive cultures subsequently sub-cultured onto solid media, including blood agar and chocolate agar, to enhance isolation of pathogens. Urine samples were plated on cystine lactose electrolyte-deficient agar and MacConkey agar to promote the growth of Escherichia coli, Klebsiella, and other common uropathogens. Body fluids were inoculated onto blood and chocolate agar. To identify respiratory pathogens, sputum samples were Gram-stained and cultured on selective media, including MacConkey and blood agar. Wound swabs were processed on blood agar and mannitol salt agar. The BD Phoenix™ M50 system was utilized for precise species-level identification and antimicrobial susceptibility testing, providing comprehensive biochemical profiles for various pathogens [10]. This combination of conventional and automated methods ensured accurate identification and susceptibility testing across all clinical sample types, adhering to CLSI (Clinical and Laboratory Standards Institute) guidelines for bacteriological analysis [11]. For samples that did not exhibit visible growth after the initial 24 hours, the incubation was extended to 48 hours.

Antibiotic susceptibility testing

Antimicrobial susceptibility testing was conducted using two methods. The classical Kirby-Bauer disk diffusion method was performed according to Clinical and Laboratory Standards Institute guidelines [11], where standardized antibiotic disks were applied to Mueller-Hinton agar plates inoculated with bacterial suspensions, and inhibition zone diameters were measured and interpreted using CLSI breakpoints (Figure 1). Additionally, the BD Phoenix™ M50 automated system was used to confirm susceptibility results and to test a broader range of antimicrobials, providing Minimum Inhibitory Concentration (MIC) values and classifying isolates as susceptible, intermediate, or resistant based on CLSI interpretive criteria. The antibiotics tested included Amikacin, Gentamicin, Gentamicin-Syn, Ampicillin-sulbactam, Ampicillin, Amoxicillin, Amoxicillin-Clavulanate, Piperacillin-Tazobactam, Piperacillin, Penicillin G, Oxacillin, Cefuroxime, Ceftriaxone, Cefepime, Cefoxitin, Ceftaroline, Cefpodoxime, Cefixime, Cefotaxime, Clarithromycin, Azithromycin, Erythromycin, Ciprofloxacin, Levofloxacin, Moxifloxacin, Norfloxacin, Ofloxacin, Trimethoprim-Sulfamethoxazole, Vancomycin, Teicoplanin, Daptomycin, Clindamycin, Tetracycline, Doxycycline, Minocycline, Tigecycline, Imipenem, Meropenem, Nitrofurantoin, Linezolid, Rifampin, Chloramphenicol, Mupirocin High level. This combined approach ensured consistent and accurate interpretation of susceptibility results, enhancing the reliability of the findings.

Antibiotic classification and multidrug resistance

The antibiotics were categorized into seven groups: aminoglycosides, beta-lactams, macrolides, sulfonamides, tetracyclines, glycopeptides, and fluoroquinolones. Multidrug-resistant (MDR) isolates were defined as bacterial strains resistant to three or more of these antibiotic classes, following established criteria [12]. This classification facilitated a comprehensive analysis of antimicrobial resistance patterns and enabled the identification of the most challenging cases of antibiotic resistance, providing critical insight into the prevalence of MDR organisms.

Data analysis

Data on bacterial isolates, antimicrobial susceptibility profiles, patient demographics, infection types, and antibiotic resistance patterns were systematically collected and entered into Microsoft Excel 2007 before being transferred to SPSS version 25 for statistical analysis. Statistical evaluations were conducted to assess differences in resistance rates between CAIs and HAIs, stratified by infection site (e.g., bloodstream, urinary tract, respiratory tract) and pathogen type. Descriptive statistics summarized the demographic and clinical characteristics of patients, while resistance rates were compared using Chi-square tests for categorical variables and t-tests for continuous variables. The analysis encompassed calculating prevalence rates, frequencies, susceptibility patterns, and other descriptive statistics, with statistical significance set at a p-value of equal to or less than 0.05 for the chi-square test, which compared categorical variables with bacterial growth.

Results

Microbial growth and participant characteristics

In this study involving 2157 participants, 1303 (60.4%) were male. Microbial growth was observed in 1177 cases (54.6%). Notably, 535 (41.1%) of the females and 445 (52.1%) of the males exhibited no growth, indicating a significant difference (p < 0.001). The mean age of participants was 43.62± 23.3years. The highest growth rate was observed in wound samples (187, 91.2%), while body fluids showed the lowest rate (171, 33.0%), reflecting a statistically significant difference (p < 0.001). The sample collection location did not significantly influence growth, with no growth in 475 (44.7%) from community settings and 216 (46.7%) from hospitals (p = 0.502).  Among the various comorbidities, obesity, renal insufficiency, and diabetes, significantly differed between participants with microbial growth and those without growth(P<0.05) (Table 1).

Distribution of isolated bacteria by setting

In this study, among the 449-gram negative bacterial isolates, 301 (67.0%) were from community settings, and 148 (33.0%) were from hospitals. Escherichia coli was the most prevalent, with 245 isolates, 179 (73.1%) from community settings and 66 (26.9%) from hospitals. Other notable gram-negative bacteria included Klebsiella pneumonia (64 isolates; 62.5% community vs. 37.5% hospital) and Pseudomonas aeruginosa (42 isolates; 50% each from community and hospital). The gram-positive bacteria primarily included Streptococcus species (100 isolates; 83(83.0%) community vs. 17(17.0%) hospital) and Enterococcus faecalis (72 isolates; 58(80.6%) community vs. 14(19.4%) hospital). Overall, gram-positive bacteria comprised 149 isolates, with a higher occurrence in community settings 284(75.3%) compared to hospitals 93(24.7%) (Table 2).

Antibiotic sensitivity and resistance in community setting

In community settings, among the tested gram-positive isolates, the highest sensitivity rates were observed for imipenem 95(96.9%), followed closely by linezolid at 151(95.6%), meropenem at 96 isolates (94.1%), tigecycline at 61 isolates (93.9%), and daptomycin at 59 isolates (93.7%). Conversely, the highest antibiotic resistance rates were recorded for azithromycin 19(90.5%), followed by ofloxacin 19 isolates (76.0%), and cefixime 68 isolates (74.7%). The overall resistance rate among gram-positive isolates was 1732 (38.7%). For gram-negative isolates, sensitivity rates were as follows: meropenem at 260 isolates (94.5%), tigecycline at 61(93.9%), imipenem at 225(85.2%), and amikacin at 128(81.5%). Notably, high resistance rates were seen, with 20 isolates (100.0%) resistant to clindamycin and 119 isolates (92.2%) resistant to ampicillin. The overall resistance rate among gram-negative isolates was 1614 (37.7%) (Suppl 1).

Antibiotic sensitivity and resistance in hospital isolates

In hospital settings, gram-positive isolates exhibited the highest sensitivity to daptomycin (43 isolates, 93.5%), followed by linezolid (62 isolates, 92.5%), teicoplanin (54 isolates, 91.5%), and tigecycline (37 isolates, 90.2%). The most significant resistance rates were observed for azithromycin (11 isolates, 84.6%) and cefixime (16 isolates, 80.0%). The overall antibiotic resistance rate among gram-positive isolates was 727 (41.1%). For gram-negative isolates, the highest sensitivity rates were noted for meropenem (108 isolates, 85.0%), imipenem (103 isolates, 79.8%), amikacin (80 isolates, 73.4%), and piperacillin-tazobactam (88 isolates, 71.5%). However, resistance was notably high for ampicillin (84 isolates, 95.6%) and cefazolin (80 isolates, 85.1%). The overall resistance rate among gram-negative isolates was 1044 (50.8%) (Suppl 2).

MDR rates in community-acquired infections

In the community setting, MDR among gram-negative bacterial isolates was observed in 183 cases (63.1%). Notably, all Morganella morganii isolates (7, 100.0%) and 3(75.0%) of Klebsiella species and Salmonella species were classified as MDR. Among gram-positive isolates, MDR was present in 171 cases (61.9%), with Lactobacillus species showing 100.0% MDR (3 isolates) and Staphylococcus aureus exhibiting a high MDR rate, with 21 out of 27 isolates (77.8%). Overall, the MDR rate in community-acquired infections was 62.5% (Suppl 3).

MDR rates in hospital-acquired infections

In the hospital setting, MDR was observed in 113 gram-negative bacterial isolates (86.2%). Notably, all isolates of Proteus species, Burkholderia cepacia, and Achromobacter species (100%) were classified as MDR. Among gram-positive isolates, 59 cases (68.6%) exhibited MDR, with Staphylococcus haemolyticus showing an MDR rate of 83.3% (10 out of 12 isolates) and Enterococcus faecalis at 78.6% (11 out of 14 isolates). Overall, the MDR rate in hospital-acquired infections was 79.3% (Suppl 3).

Risk factors for community vs. hospital-acquired infections

In the analysis of risk factors for community-acquired versus hospital-acquired infections, males had a significantly higher proportion of hospital-acquired infections, with 688 (75.4%) compared to 374 (61.0%) in community-acquired infections (p<0.001). Individuals over 40 years old were more likely to have hospital-acquired infections, 280 (35.2%) versus 183(25.1%) in the community-acquired group (p<0.001). Patients with pre-existing comorbidities, including diabetes, malignancy, obesity, hypertension, renal insufficiency, heart failure, and asthma, demonstrated significantly higher rates of hospital-acquired infections (p < 0.05) (Table 3).

Discussion

Antimicrobial resistance (AMR) has become one of the most critical global public health challenges of the 21^st century. It arises when microorganisms resist antimicrobial drugs such as antibiotics, rendering these treatments ineffective. This resistance primarily results from the overuse and misuse of antibiotics in various sectors, including clinical settings. Often referred to as the "Silent Pandemic," AMR demands immediate and effective action rather than being treated as a distant concern [13]. Despite the growing threat of antimicrobial resistance, the overuse of these agents remains prevalent, particularly in patients with critical illnesses, advanced disease stages, malignancies, or immunocompromised conditions [14].

Hospitals are recognized as high-risk environments for health, particularly due to the prevalence of HAIs in both developed and developing countries [15]. The impact of HAIs is substantial, contributing to increased healthcare costs, greater disease severity, higher rates of antimicrobial resistance, and elevated morbidity and mortality. Within healthcare settings, bacterial pathogens are the primary culprits behind nosocomial infections, with many strains exhibiting resistance to both standard and last-resort antibiotics [16].

Gram-negative bacteria are frequently involved in HAIs, accounting for up to 87% of cases [15]. Among Gram-positive bacteria, Staphylococcus aureus is the most prevalent strain [17]. In Europe and Asia, the most common Gram-negative pathogens include Pseudomonas aeruginosa, Acinetobacter baumannii, and members of the Enterobacteriaceae family [18,19]. A multicenter retrospective study conducted across five private hospitals in Lebanon, involving 258 patients, reported that Escherichia coli and Pseudomonas aeruginosa were the most prevalent Gram-negative bacteria, while Staphylococcus aureus was the dominant Gram-positive isolate [1]. Similarly, the present study found that Gram-negative bacteria accounted for 62.1% (148 out of 241) of hospital-acquired infections (HAIs). The most frequently isolated Gram-negative pathogens were Escherichia coli, Klebsiella pneumoniae, and Pseudomonas aeruginosa. Among Gram-positive bacteria, Staphylococcus aureus emerged as the most prevalent strain in the hospital setting.

Hospital-acquired microorganisms exhibited greater resistance to antimicrobials than their community-acquired counterparts. For instance, a study by Matta et al. demonstrated that hospital-acquired Pseudomonas aeruginosa showed significantly higher resistance to all evaluated antimicrobial agents thanacquired strains [1]. In the current study, the resistance rate among community-acquired infections was 38.2% (3,346 out of 8,760 isolates), whereas the resistance rate among hospital-acquired infections was 46.3% (1,771 out of 3,825 isolates).

Escherichia coli infections are typically treated with antibiotics such as ciprofloxacin, levofloxacin, and other fluoroquinolones; however, resistance to multiple antibiotics has become increasingly prevalent. The growing resistance to fluoroquinolones and the emergence of extended-spectrum beta-lactamases pose significant challenges in managing these infections. Although carbapenems are generally considered the preferred treatment for MDR Escherichia coli infections, reports of resistance to carbapenems are also rising [15]. In this study, sensitivity rates for Escherichia coli isolates in community-acquired infections were found to be 49.4%, 53%, and 60.3% for ciprofloxacin, levofloxacin, and norfloxacin, respectively. In contrast, sensitivity rates among hospital-acquired isolates were lower, with 29.7%, 33.3%, and 28.6% for the same antibiotics. Furthermore, sensitivity to imipenem and meropenem was observed in 91.9% and 96.0% of community-acquired Escherichia coli isolates, while sensitivity in hospital-acquired cases was notably lower at 83.3% and 87.9%. These findings indicate a concerning trend of increased antibiotic resistance among Escherichia coli isolates from hospital settings, particularly concerning carbapenem resistance.

Klebsiella pneumoniae is the second most prevalent cause of HAIs, following Escherichia coli [15]. While it is primarily considered an opportunistic pathogen, there has been a notable increase in its hypervirulence, often linked to hypercapsulation [20], along with a rise in antibiotic resistance [21]. The emergence of carbapenem-resistant Klebsiella pneumoniae strains poses a significant global health threat, contributing to increased mortality rates primarily due to the acquisition of Klebsiella pneumoniae carbapenemases [22]. Multidrug-resistant strains can exhibit resistance to all beta-lactams and fluoroquinolones. Consequently, last-resort treatment options often involve polymyxin B, frequently in combination with tigecycline or certain aminoglycosides [15]. In this study, community-acquired Klebsiella pneumoniae isolates showed a sensitivity rate of 100% to tigecycline, whereas the sensitivity among hospital-acquired isolates was significantly lower at 68.4%. Additionally, fewer than 50% of K. pneumoniae isolates demonstrated sensitivity to all beta-lactam antibiotics.

A study conducted in India investigating the etiology and antimicrobial sensitivity of organisms responsible for community-acquired pneumonia, which included 145 patients, found Streptococcus infections to be the most frequently isolated pathogen in the community setting [23]. In line with these findings, the current study also identified Streptococcus infections as one of the most commonly isolated pathogens within the community context. This could be explained by high transmissibility, opportunistic nature in vulnerable populations, association with diverse infections, seasonal peaks, and the dynamics of antimicrobial resistance and vaccination.

In recent decades, the prevalence of antimicrobial resistance has escalated worldwide, with MDR bacteria emerging as a significant cause of nosocomial infections. The risk of MDR infections is linked to several factors, including prolonged antimicrobial therapy, cross-transmission, extended hospital stays, and invasive procedures. These resistant bacteria can lead to various infections—such as pneumonia, urinary tract infections, and wound infections—associated with increased morbidity, and mortality [24]. In this study, a higher MDR was found among HAIs compared to Community acquired infections, with 79.3% for HAI and 62.5% for Community acquired setting.  Higher rates of MDR in HAIs compared to community-acquired infections result from factors such as prolonged antibiotic use, invasive procedures, and close patient proximity, which foster the emergence and spread of resistant strains. A retrospective study conducted in a tertiary general hospital in Jining, China, revealed a high prevalence of MDR HAIs; out of 7,579 bacterial isolates, 3,223 (42.5%) were identified as MDR. Gram-negative bacteria were the most frequently isolated MDR pathogens, with Escherichia coli exhibiting the highest detection rate at 37.7%. Collectively, Escherichia coli and Klebsiella pneumoniae accounted for 51.0% of all MDR isolates [24]. In this study, the prevalence of MDR among hospital settings was found to be 79.3%, with 172 out of 217 isolates classified as MDR. Consistent with previous findings, gram-negative bacteria were the most frequently isolated MDR pathogens, with Escherichia coli detected in 24.9% (54 out of 217) of cases, followed by Klebsiella pneumoniae at 10.1% (22 out of 217).

Multidrug-resistant Staphylococcus aureus is a leading cause of HAIs and a significant contributor to mortality among hospitalized patients, largely due to its possession of resistance genes against various antibiotics, including commonly used anti-staphylococcal drugs [25]. In this study, 25.4% (15 out of 59) of the MDR Gram-positive isolates from hospital settings were identified as multidrug-resistant Staphylococcus aureus. Most Staphylococcus aureus isolates exhibited resistance to penicillin, while all were sensitive to the carbapenems.

A prospective cohort study conducted over one year at a university tertiary care hospital in Portugal identified neoplastic diseases, including hematologic malignancies and solid tumors, as well as immunocompromised states, as common conditions associated with hospital-acquired infections [26]. Notably, no gender differences were observed in infection rates [1]. In this study, patients with pre-existing comorbidities such as diabetes, malignancy, obesity, hypertension, renal insufficiency, heart failure, and asthma were found to have significantly higher rates of hospital-acquired infections.

Conclusion

Multidrug-resistant infections were prevalent in HAIs, with most isolates resistant to current antibiotics. This underscores the need for enhanced surveillance to optimize antibiotic use and control HAIs. The higher resistance in HAIs compared to community-acquired infections highlights the importance of early detection of resistance.

Declarations

Conflicts of interest: The author(s) have no conflicts of interest to disclose.

Ethical approval: The study's ethical approval was obtained from the scientific committee of the Kscien Organization for Scientific Research.

Patient consent (participation and publication): Verbal informed consent was obtained from patients for participation in this study and publication.

Source of Funding: Saaeda company.

Role of Funder: The funder remained independent, refraining from involvement in data collection, analysis, or result formulation, ensuring unbiased research free from external influence.

Acknowledgements: None to be declared.

Authors' contributions: BAA and FHK were significant contributors to the conception of the study and the literature search for related studies.  RQS, HAY, WAH and SHK were involved in the literature review, the study's design, and the critical revision of the manuscript, and they participated in data collection. AMM, FA, and KKM were involved in the literature review, study design, and manuscript writing. DQM, BHI, HSA, SHA, MOS and SSA Literature review, final approval of the manuscript, and processing of the tables. RQS and AMM confirm the authenticity of all the raw data. All authors approved the final version of the manuscript.

Use of AI: AI was not used in the drafting of the manuscript, the production of graphical elements, or the collection and analysis of data.

Data availability statement: Not applicable.

Abstract

Introduction

There has been a notable rise in antibiotic resistance among enterobacteria. This issue is primarily attributed to the emergence of extended-spectrum beta-lactamases (ESBLs), which present a significant concern for public health worldwide. This study investigates the prevalence of ESBL production, antibiotic resistance profiles, and molecular identification of the blaTEM gene in Klebsiella pneumoniae isolates.

Methods

The samples were randomly collected from several medical facilities in Erbil city. The VITEK 2 system was used for bacterial identification, antibiotic susceptibility, and ESBL production testing. The Double Disc Synergy Test (DDST) confirmed ESBL production. Polymerase chain reaction was conducted on all DNA samples, and the amplified DNA was analyzed using agarose gel electrophoresis to detect the blaTEM gene.

Results

A total of 43 samples were collected, of which the majority were urine (56%), followed by sputum (28%), blood (9%), and wound (7%). Klebsiella pneumoniae isolates exhibited the highest prevalence of resistance against ceftazidime (72%), ceftriaxone (70%), ciprofloxacin (63%), amoxicillin-clavulanic acid (60%), amikacin (58%), cefotaxime/tazobactam (56%), and gentamicin (53%). The DDST results indicated positive ESBL production in 15 isolates (35%), as evidenced by an increase or distortion in the inhibition zone toward the amoxicillin-clavulanate disc. Of the 43 isolates, 34 (79%) carried the blaTEM gene.

Conclusion

The study area shows a significant level of antibiotic resistance in ESBL-producing Klebsiella pneumoniae isolates, which, if not adequately addressed, could soon lead to severe health and therapeutic consequences.

Introduction

Antibiotic resistance is a dynamic and ongoing challenge. Over the past three decades, there has been a notable rise in antibiotic resistance among enterobacteria, particularly concerning third-generation cephalosporins. This issue is primarily attributed to the emergence of extended-spectrum beta-lactamases (ESBLs). ESBLs are a group of beta-lactamase enzymes produced by gram-negative bacteria that can hydrolyze and inactivate many antibiotics, including aztreonam, penicillin, and cephalosporins. These enzymes, such as TEM, SHV, and CTX-M, and their variants, allow enterobacteria to resist β-lactam antibiotics. The encroachment of these multi-drug-resistant pathogens into community settings raises significant alarm. The swift spread of the genes that encode these ESBLs, primarily via plasmids, has facilitated their rapid proliferation, leading to an alarming increase in the global prevalence of ESBL-producing bacteria. This situation presents a significant concern for public health worldwide [1,2]. Klebsiella pneumoniae is an opportunistic pathogen responsible for nosocomial and community-acquired infections [3]. This bacterium is usually found as an intestinal microflora. It can cause severe diseases like ventilator-associated pneumonia, catheter-related urinary tract infections, meningitis, bacteremia, septicemia, infection of surgical and non-surgical wounds, diarrhea, prosthetic valve endocarditis, peritonitis, and osteomyelitis [1]. The first strains of Klebsiella pneumoniae with ESBL (KP-ESBL) were identified in Europe in 1982, marking the emergence of resistance to ceftazidime and aztreonam due to plasmid-mediated beta-lactamase enzymes. These resistance traits rapidly disseminated to other gram-negative bacteria, including Escherichia coli. Since the discovery of these enzymes, their prevalence has continued to rise, with over 200 distinct ESBL enzymes identified today. The impact of ESBL-producing strains is particularly pronounced in intensive care units, which pose a high risk for epidemic outbreaks. K. pneumoniae and E. coli are the two most frequently encountered bacterial species associated with ESBL production. Implementing proactive monitoring of ESBL-producing pathogens in high-risk populations is crucial through suitable antimicrobial strategies. This necessity arises from the tendency of these pathogens to display multidrug resistance, which complicates treatment options and increases the potential for severe infections [4]. This study investigates the prevalence of ESBL production, antibiotic resistance profiles, and molecular identification of the blaTEM gene in Klebsiella pneumoniae isolates in Erbil City.

Methods

Study design and sample collection

This retrospective cross-sectional study was conducted between September 2021 and January 2022. The ethics board of the College of Health Sciences at Hewler Medical University approved the study. Patient consent was verbally obtained from the patients to collect their samples and be published in this study.  The samples were randomly collected from several medical facilities in Erbil city, such as Nanakali Hospital, Hawler Teaching Hospital, and private laboratories. The collected samples included urine, sputum, or wound infection.

Isolation, identification, and susceptibility test of Klebsiella pneumoniae

The samples were cultured on MacConkey agar and incubated at 37°C for 24 hours. Klebsiella pneumoniae colonies were isolated based on their large, pink-to-red mucoid appearance. The VITEK 2 system (bioMérieux, France) was used for bacterial identification and antibiotic susceptibility testing. A sterile swab or stick was used to transfer an adequate number of colonies from the cultured sample into 3.0 ml of sterile saline (0.45% to 0.50% NaCl, pH 4.5 to 7) in a 12 × 75 mm polystyrene test tube. The suspension’s turbidity was then adjusted to match the 0.5 McFarland standard for antimicrobial susceptibility testing and measured using a DensiChek™ turbidimeter.

Phenotypic Detection and Confirmation of ESBLs

ESBL production was detected among the isolates using the VITEK 2 system. The testing panel included six wells containing 10 μg/ml ceftriaxone, 0.5 μg/ml cefotaxime, or 0.5 μg/ml ceftazidime, alone or in combination with clavulanic acid (10 μg/ml, 4 μg/ml, and 4 μg/ml, respectively). Bacterial growth in each well was quantitatively assessed using an optical scanner. A proportional reduction in growth in the wells containing cephalosporin combined with clavulanic acid, compared to those containing cephalosporin alone, was considered indicative of ESBL production. The Double Disc Synergy Test (DDST) confirmed ESBL production. Each Klebsiella pneumoniae isolate was inoculated on Müller-Hinton agar plates for susceptibility testing. Discs containing 30 μg of cefotaxime and ceftazidime were placed on either side of a disc with co-amoxiclav (20/10 μg), positioned 20 mm apart (center to center). ESBL production was indicated when the inhibition zone of either cephalosporin was expanded by clavulanic acid, often resulting in a distinctive "champagne-cork" or "keyhole" shape zone [1].

Molecular detection of blaTEM gene

DNA extraction of Klebsiella pneumoniae was carried out using the simple boiling method. Isolated colonies from overnight cultures were placed in a test tube containing 1 ml of distilled water and heated in a water bath for 10 minutes. The sample was centrifuged at 1,000 rpm for 5 minutes to complete the extraction process. Polymerase chain reaction (PCR) was conducted on all DNA samples of Klebsiella pneumoniae. The concentration of extracted double-stranded DNA was combined with a PCR master mix (Promega, USA), which included Taq DNA polymerase, deoxynucleotide triphosphates (dATP, dCTP, dGTP, dTTP), buffers, and salts. Specific forward (5'-GAG TAT CAA CAT TTC CGT GTC-3') and reverse (5'-TAA TCA GTG GGC ACC TTC TC-3') primers were added to amplify the blaTEM target gene. The PCR mixture was then incubated in a thermocycler, undergoing 32 cycles of alternating temperatures: a denaturation step at 94°C for one minute to separate the double-stranded DNA into single strands, an annealing step at 57°C for one minute to allow the primers to bind to the target DNA, and an extension step at 72°C for 10 minutes for the Taq DNA polymerase to synthesize new DNA strands. This process generated billions of copies of the target DNA sequence within a few hours. Subsequently, the amplified DNA was separated and analyzed using agarose gel electrophoresis.

Results

A total of 43 samples were collected; 22 (51%) were obtained from females and 21 (49%) from males. The majority of the samples were urine (56%), followed by sputum (28%), blood (9%), and wound (7%). Klebsiella pneumoniae isolates exhibited the highest prevalence of resistance against ceftazidime (72%), ceftriaxone (70%), ciprofloxacin (63%), amoxicillin-clavulanic acid (60%), amikacin (58%), cefotaxime/tazobactam (56%), and gentamicin (53%). In contrast, the lowest prevalence of resistance was observed with imipenem (26%) and meropenem (28%) (Table 1). The DDST results indicated positive ESBL production in 15 isolates (35%), as evidenced by an increase or distortion in the inhibition zone toward the amoxicillin-clavulanate disc. The remaining samples (65%) tested negative for ESBL production (Figure 1). Of the 43 isolates, 34 (79%) carried the blaTEM gene (Figure 2). Overall, 11 isolates (26%) tested positive in the DDST and PCR, while five isolates (12%) were negative. Additionally, four isolates (9%) were positive in DDST but negative in PCR, whereas 23 isolates (53%) were negative in DDST but positive in PCR (Table 2). 

Discussion

KP-ESBL has emerged as a significant concern due to its virulence factors and role as a leading cause of infectious diseases. It is categorized within the multidrug-resistant group of bacteria. Klebsiella pneumoniae exhibits antibiotic resistance through several mechanisms, including the enzymatic degradation or inactivation of antibiotic compounds, alterations in membrane permeability, and modifications of antibiotic target sites via bacterial protein mutations. It has been reported that K. pneumoniae acquires numerous ESBL enzymes, which serve as its primary defense mechanism against antibiotics [5]. The prompt and precise identification of ESBL-positive Enterobacteriaceae strains is critical for guiding appropriate patient management and developing and enforcing targeted infection control protocols [6]. Over the past decade, numerous studies have highlighted the prevalence of ESBL-mediated resistance in infectious bacteria, with a higher prevalence in Escherichia coli and Klebsiella pneumoniae [2,5,6-9]. In the study by Benbrahim et al., 15 out of 40 (37.5%) Klebsiella pneumoniae isolates were identified as ESBL-producing strains, a figure close to the 41.1% reported by Pirzaman et al. [10]. Another study reported a prevalence of 31.8% [2]. Higher rates were observed in Asian countries, reaching up to 75% [1]. In this study, the prevalence of isolates with positive ESBL enzymes was 35%, close to previously reported [1,2,10].

Infections caused by ESBL-producing bacteria can affect individuals of all ages, though their distribution may be influenced by patients' immunological status and the prevalence of antibiotic misuse. Benbrahim et al. found that KP-ESBLs were detected across all age groups, with the highest incidence (20%) observed in individuals aged 21-30 [1].

Gravey et al. reported a KP-ESBL prevalence of 4.1% in individuals aged 18–64 and 4.2% in those over 65 [11]. Due to the retrospective nature of the present study, the age preference among samples was unknown. In the present study, females comprised 51% of the sample, while males accounted for 49%, a distribution that closely aligns with the findings reported by Marra et al. [12]. However, a study by Ali and Ismael observed a higher proportion of female samples (72.72%) than male samples (22.73%) among 84 Klebsiella pneumoniae isolates [13]. Conversely, a study by Nirwati et al. found a higher frequency of male samples (64.07%) than female samples (35.93%) in 167 K. pneumoniae isolates [14].  In the study by Benbrahim et al., KP-ESBL strains were most frequently isolated from urine samples, which the authors suggested is likely due to urine being one of the most commonly collected specimens for clinical investigation [1]. Similarly, most of the samples (56%) in the present study were urine.

Antimicrobial resistance in pathogenic bacteria presents a global challenge, contributing to high mortality and morbidity rates. Infections caused by multidrug-resistant strains are often difficult or impossible to treat with conventional antimicrobials. The widespread, usually unnecessary use of antibiotics stems from the failure of many healthcare centers to promptly diagnose causative microorganisms and determine their antimicrobial susceptibility in patients with bacteremia and other serious infections [15]. In this study, Klebsiella pneumoniae strains exhibited the highest resistance rates against ceftazidime (72%), ceftriaxone (70%), ciprofloxacin (63%), and amoxicillin-clavulanic acid (60%). The resistance rate for ceftazidime in this study was slightly similar to that reported in another study conducted in Erbil, which documented a resistance rate of 62.5% [16]. Notably, our study found significantly higher resistance rates for ciprofloxacin (63% vs. 4.2%), gentamicin (53% vs. 4.2%), and amikacin (58% vs. 4.2%). Conversely, the resistance rate for amoxicillin-clavulanic acid was lower in our study than in the previous one (60% vs. 81.25%) [16]. Another study conducted in Iran reported even higher resistance rates for amoxicillin-clavulanic acid (100%), ceftazidime (84%), tazobactam (80%), ciprofloxacin (80%), and amikacin (60%) [17]. In contrast to our findings, sensitivity rates of 90% and 100% by K. pneumoniae isolates have been reported for amoxicillin-clavulanic acid and imipenem, respectively [1,18]. Muggeo et al. and Rasamiravaka et al. found a sensitivity rate of 83% to amikacin, while it was 42% in the present study. In the study by Bora et al., imipenem and meropenem were identified as one of the most effective antimicrobial agents against ESBL-producing isolates, comparable to the findings in this study [7].

The prevalence of the blaTEM gene in Klebsiella pneumoniae isolates varies across studies. Alibi et al. reported that blaTEM gene was present in 56.8% of their isolates [19]. In contrast, Paterson et al. found a prevalence of 87% [20]. Guessennd et al. reported that 63.4% of their isolates in Abidjan, Côte d’Ivoire, carried the blaTEM gene [21]. The prevalence rates of 31.57% and 55% for the blaTEM gene in Klebsiella pneumoniae have also been reported [2,5]. Based on genotypic detection, in this study, 79% of the isolates carried the blaTEM gene, which may indicate uncontrolled antibiotic misuse in our locality. In line with this finding, Bora et al. found that genes responsible for ESBL were positive in 79.45% of the Klebsiella pneumoniae isolates. However, in the phenotypic confirmatory test for ESBL production, 53.42% of the isolates tested positive, suggesting that the phenotypic test results may be subject to false positives or negatives. Several factors can contribute to this issue, including the production of multiple β-lactamase types by a single bacterial isolate, the co-production of ESBLs and constitutive AmpC β-lactamases, variations in substrate affinities, and the inoculum effect [7]. Accordingly, in the current study, phenotypic detection revealed that only 15 isolates (34.9%) were positive for ESBL production, while genotypic detection identified 34 isolates (79%) as positive for the blaTEM gene. Therefore, PCR using oligonucleotide primers specific to ESBL genes is considered the most straightforward and reliable method for detecting the presence of ESBLs [7].

This study has several limitations that may impact the generalizability of its findings, including small sample size, sampling restricted to a specific geographic area of the city, and the exclusive focus on the blaTEM gene. Future research with larger sample sizes and broader geographic coverage, extending beyond the city to the entire country, is needed to provide deeper insights into this issue.

Conclusion

The study area shows a significant level of antibiotic resistance in KP-ESBL strains, which, if not adequately addressed, could soon lead to severe health and therapeutic consequences. Molecular methods, which detect the genes responsible for resistance, may offer greater sensitivity in identifying antibiotic resistance.

Declarations

Conflicts of interest: The author(s) have no conflicts of interest to disclose.

Ethical approval: The ethics board of the College of Health Sciences at Hewler Medical University approved the study.

Patient consent (participation and publication): Written informed consent was obtained from patients for publication.

Source of Funding: Medical Research Center of Hawler Medical University.

Role of Funder: The funder remained independent, refraining from involvement in data collection, analysis, or result formulation, ensuring unbiased research free from external influence.

Acknowledgements: None to be declared.

Authors' contributions: MAG and KAO were significant contributors to the conception of the study and the literature search for related studies.  LJM, LKHJ, and KAO were involved in the literature review, the study's design, and the critical revision of the manuscript, and they participated in data collection. HOA and KWM were involved in the literature review, study design, and manuscript writing. SYI and BMI Literature review, final approval of the manuscript, and processing of the tables. HOA and MAG confirm the authenticity of all the raw data. All authors approved the final version of the manuscript.

Use of AI: AI was not used in the drafting of the manuscript, the production of graphical elements, or the collection and analysis of data.

Data availability statement: Not applicable.

Abstract 

Introduction

Transfusion medicine promotes the safety of blood transfusions by rigorously testing to eliminate risks of infection and hemolytic. The efficacy (to correct and identify antibodies) of a modified cross-matching method that excludes the use of anti-human globulin (AHG) bovine serum with that of the immediate spin crossmatch was assessed.

Material and Methods

This multi-center study was performed at two medical centers in Iran and Iraq. Over seven years, consecutive participants received two different blood cross-matching methods: one with AHG and bovine serum and another without it.

Results

The study included 31240 participants. About 18526 (59.3%) were males and 12714 (40.7%) were females. The ages of participants ranged from 23 to 57 years, with the average age of 39 years. Only 45 (0.14%) participants blood cross-match incompatible in both spin cross-match and cross-match without AHG, bovine serum the same result. The study found that all detected antibodies correlated with potential blood incompatibility, and there remained 100% safety between the two testing methods.

Conclusion

Omitting AHG and bovine serum in cross-matching might be safe. At least it can be used in emergency situations and in resource-limited settings. 

Introduction

A transfusion medic is in charge of making sure the blood you are being transfused is being transfused as safely as possible. To address concerns such as transfusion-transmitted infections (TTIs) and transfusion-associated hemolytics [1], an exhaustive testing procedure is performed between the donor and recipient blood. The present red cell serology based on safe transfusion was developed in 1901 with Karl Landsteiner's identification of the ABO blood types. Later, in that same year, 1940, the Rh antigen and in the year 1946, the Kell antigen were shown to be significant red cell antigens [2]. Provided donor and recipient blood types are compatible, these minimally increase the risks of medical procedure complications, adverse reactions, and other conditions that benefit from medical procedures such as organ transplants [3,4]. Cross-matching aims to see if the blood types from the donor and receiver are compatible [5]. This is done to avoid potentially fatal immunological responses to transfused blood. This approach can search the recipient's blood for antibodies that are incompatible with the red blood cell antigens in the donor’s cells [6,7]. IgM or IgG are usually blood group antibodies, but occasionally IgA. In most cases, immune antibodies (generally IgG) and naturally occurring antibodies (predominantly IgM) [8]. The Anti human Globulin (AHG) and bovine serum detect weak antibodies, such as most IgG and some IgM antibodies. However, the bovine serum also detects weak antibodies such as IgG due to increasing its dielectric constant and hence decreasing the Zeta potential, which causes the red cell to come closer, thereby decreasing the thickness of the ionic cloud surrounding each cell, resulting in agglutination [9]. AHG helps reveal the presence of antibodies that may have coated the surface of red blood cells even when agglutination is not apparent during the initial testing phases; adding AHG and bovine to the test significantly increases the test's sensitivity [10]. These antibodies may not cause visible clumping or agglutination on their own; however, not all antigens lead to the formation of clinically significant antibodies to identify the 25–28 blood group antigens that are known to elicit hemolytic reactions (HTRs), screening for significant antibodies in the blood cells is recommended [11]. Antibody screening tests are of significant use in both the diagnosis of infectious diseases and the testing of immune systems, as well as the discovery of autoimmune conditions that can be gained from the use of antibody screening tests. As part of this procedure, the majority of blood banks routinely perform a battery of tests to identify the ABO and Rh types of both the donor and the recipient and then carry out a full cross-match of the recipient’s serum with the donor’s red blood cells to ensure transfusion stability, at times being an integral part of this procedure is a blood cross-match with bovine serum plus AHG.

The objective of the study was to assess the possibility of making the blood cross machining method safe and accurate, when AHG and bovine serum were not utilized in the process.

Material and Methods

This multi-center cohort study was conducted at the Blood Bank Department of General Hospital in Kurdistan, Iraq, and Kawsar Hospital in Sanandaj, Iran. The research spanned seven years (from December 2017 to November 2024). All patients necessitated blood transfusions, and some volunteers participated in the spin cross-match (blood cross-match with AHG, bovine) [13]. All tests were repeated blood cross-matches without AHG, bovine (Figure 1). (Inclusion criteria: patients and volunteers must possess identical ABO and Rh blood types.

Only 45 participants (1.4%) exhibited blood cross-match incompatibility in both spin cross-match and cross-match without AHG and bovine; the same result was observed where they tested positive for antibodies. Antibody screening for incompatible patients was conducted using the Bio-Rad ZE5 Cell Analyzer, a high-performance flow cytometry platform designed for a diverse array of cellular assays, providing advanced capabilities for researchers in immunology and oncology. It combines flexibility with high throughput. Data analysis each variable underwent a descriptive analysis, calculating percentages, frequencies, means, and ranges. The data were gathered on an Excel sheet. Whenever necessary, the P-value significance was set to less than 0.05.

Results

The study included 31240 participants. About 18526 (59.3%) were males and 12714 (40.7%) were females. Participants’ ages ranged from 23 to 57 years, with an average age of 39. Only 45 (1.4%) participants' blood cross was incompatible in spin cross-match and cross-match without AHG; bovine serum. From this number, 27 (53.3%) had a history of blood transfusion, 16 (59% male), and 11 (41% female). About 18 (46.7%) of women who have been pregnant at least once are included [table 1].

The results of the two tests about compatibility were identical in every way [Figure 2 and 3].  No discrepancy was found between the two procedures.

Discussion

The International Society of Blood Transfusion has recognized 33 blood group systems, and that's a significant step forward. Other than the ABO and Rhesus systems, red blood cell membranes contain a plethora of different antigens. This is important to avoid transfusion-related problems [14]. One of the common hazards associated with red blood cell transfusions is the formation of RBC alloantibodies. Red blood cell alloimmunisation rates are notable in 1% to 35% of some groups of patients [15,16]. In this study, the alloimmunisation rate was far lower at 0.14%. The distribution of alloantibodies was determined in 35.5% (16 out of 45) of the cases where anti-D was found. About 24.4% (11 out of 45) were anti-M. The third most common group was Anti-C antibodies, at about 20%.

In Western nations, studies have shown a much higher incidence of Kell antibodies [16], which contradicts our results. This might need further studies with a broader population. The striking finding of this study was that there was not a single case where the spin cross was compatible, but without AHG, the bovine cross-match was incompatible, which is an important finding. The screening cells found every clinically relevant antibody. After the cross-match method, it was found that this screening panel is suitable for use in Iranian and Iraqi blood banks, at least in special situations like emergency settings or resource-limited regions. 

This demonstrates that if the AHG bovine had been deleted from the blood and the blood was supplied using this approach, an identical result would have been observed using the AHG bovine in vitro. Identifying all clinically important antibodies is a crucial part of the pre-transfusion testing process and helps avoid an adverse response. The findings of the test are interpreted according to the presence or lack of agglutination or other markers of antigen binding [17]. 

On the other hand, eliminating the traditional AHG, bovine serum cross-match, may be advantageous.  There is a reduction in the amount of labor that has to be done, a reduction in the prices of the reagents, a straightforward and speedy method, suitableness for emergencies, sensitivity and accuracy, fast detection of incompatible conditions, and more efficient utilization of blood inventory.

The technical staff will not have to carry out an AHG, bovine serum cross-match every ten to fifteen minutes. As a result, they will have more time to dedicate to other regions, such as donor recruiting and well-being. This cross-match without AHG bovine serum is also excellent for the development of a blood bank. A procedure for computer cross-match might potentially be created with the confidence obtained from this work in the future, provided that the software and checking points that are used are confirmed. However, confirming these benefits needs several better-designed studies with more variables.

This study has several limitations: First, although the sample size is reasonable, there is no significant data regarding the cost of the procedures in both situations. Second. The duration of both tests has not been reported.  Third, we failed to report the number of agglutinations found on each test plate.

Conclusion

Based on this study's findings, crossmatching without the use of AHG and bovine serum demonstrated a predicted safety rate of 100%, yielding results comparable to those of the spin crossmatch method. This suggests that omitting AHG and bovine serum in crossmatching may be a safe alternative, particularly in emergencies and resource-limited settings, where simplifying procedures can help optimize care without compromising safety.

Declarations

Conflicts of interest: The author(s) have no conflicts of interest to disclose.

Ethical approval: The ethics board of the Kurdistan University of Medical Science approved the study.

Patient consent (participation and publication): Written informed consent was obtained from patients for publication.

Source of Funding: Kurdistan University of Medical Science.

Role of Funder: The funder remained independent, refraining from involvement in data collection, analysis, or result formulation, ensuring unbiased research free from external influence.

Acknowledgements: None to be declared.

Authors' contributions: MRR significantly contributed to the study's conception and literature search for related studies. SHK and MB were involved in the literature review, the study's design, and the critical revision of the manuscript. They also participated in data collection, the literature review, study design, and manuscript writing. MRR and SHK confirmed the authenticity of all the raw data. All authors approved the final version of the manuscript.

Use of AI: AI was not used in the drafting of the manuscript, the production of graphical elements, or the collection and analysis of data.

Data availability statement: Not applicable.