Predictors of post-operative complications following thyroid surgeries at a tertiary hospital in Tanzania

Abstract

Background

Thyroid disorders pose a substantial burden in low- and middle-income countries (LMICs), where late presentation and limited access to specialized care often result in advanced disease at surgery. Despite thyroidectomy being the definitive treatment for many thyroid conditions, data on surgical outcomes and complication rates from sub-Saharan Africa remain scarce. This study aimed to describe the clinical characteristics, surgical approaches, and in-hospital postoperative outcomes of patients undergoing thyroidectomy at a tertiary hospital in Tanzania, and to identify factors associated with postoperative complications.

Materials and Methods

This was a hospital-based retrospective cohort study involving patients who underwent thyroid surgery between January 2018 and December 2023. Data were retrieved from case notes and surgical records in the hospital’s medical records department. Sociodemographic and clinical characteristics were analyzed as predictor variables, while postoperative in-hospital complications were the primary outcomes. Multivariable logistic regression was performed to identify independent predictors of complications.

Results

A total of 405 patients underwent thyroidectomy during the study period, with a female predominance (89.6%) and a mean age of 48 ± 13 years. The majority presented with anterior neck swelling (98.5%), and 24% had retrosternal extension. Benign pathology accounted for 77% of cases. The overall complication rate was 15.3%, with hypocalcemia (4%) and superior laryngeal nerve injury (3%) being the most common. The mortality rate was 0.5%. Independent predictors of in-hospital complications included total thyroidectomy (aOR: 3.12; 95% CI: 1.08–9.02) and symptom duration exceeding 10 years (aOR: 4.12; 95% CI: 1.60–10.58).

Conclusion

Postoperative morbidity following thyroidectomy in this setting was primarily driven by delayed presentation and the extent of surgery. Targeted interventions to promote early diagnosis and optimize perioperative care are essential to improve patient outcomes.

Introduction

Thyroid disorders are among the most common endocrine diseases globally, second only to diabetes, and affect millions of individuals across all geographic and socioeconomic backgrounds [1]. These disorders range from benign goiters and thyroid nodules to malignant lesions, often presenting with symptoms such as anterior neck swelling, dysphagia, dyspnea, or signs of hyper- or hypothyroidism. The burden of disease is particularly pronounced in low- and middle-income countries (LMICs), where access to diagnostic and therapeutic services is often limited [2]. In such settings, delayed presentation is common, with patients frequently seeking care only when symptoms become severe or disabling [3]. This delay not only complicates the clinical picture but also increases the likelihood of encountering large or retrosternal goiters, invasive malignancies, or significant comorbidities that increase the surgical risk [4].

Thyroidectomy remains the cornerstone of management for a variety of thyroid diseases, including multinodular goiter, differentiated thyroid carcinoma, and compressive or symptomatic lesions [5,6]. In well-resourced environments, thyroidectomy is considered a routine and safe procedure, with complication rates typically below 5 % in high-volume centers [[7], [8], [9]]. However, these favorable outcomes may not be directly generalizable to LMICs, where systemic limitations in surgical infrastructure, preoperative evaluation, intraoperative monitoring, and postoperative care may adversely affect surgical outcomes [10]. Despite the high prevalence of thyroid disease in sub-Saharan Africa [2], there is a paucity of data on the clinical presentation, surgical patterns, and postoperative outcomes of thyroidectomy patients in the region.

To address these gaps, this study aimed to describe the demographic and clinical characteristics of patients undergoing thyroidectomy at a high-volume quaternary hospital in Tanzania and to determine the magnitude, severity, and predictors of post-thyroidectomy in-hospital complications. Findings from this study are intended to inform clinical decision-making, improve risk stratification, and support the development of context-appropriate surgical protocols and capacity-building strategies in LMIC settings.

Methods

Study design and setting

This was a retrospective cohort study involving patients who underwent thyroidectomy surgeries at Muhimbili National Hospital (MNH) between January 2018 and December 2023. MNH is the highest referral level in the Tanzanian health system. It serves as the teaching facility for Muhimbili University of Health and Allied Sciences (MUHAS) and other institutions countrywide.

Study population and sample

To be included in the study, patients had to be 18 years or older, have undergone thyroidectomy at MNH between January 2018 and December 2023, and have a confirmed post-thyroidectomy histology result. We excluded patients with no histological results.

A post hoc power analysis based on the sample of 405 patients indicated that the study had 92 % power to detect the observed proportion of patients who developed complications following thyroidectomy, at a two-tailed significance level of α = 0.05.

Study participants and data collection

We identified patients who underwent thyroidectomy at MNH between January 2018 and December 2023 using the Electronic Medical Record System (EMS) theatre booking lists, which were cross-referenced with the theatre logbooks. Using the identified patients' hospital numbers, individual case notes were retrieved from the medical records department and the EMS. Variables were abstracted using a pretested abstraction tool. Independent variables included the patient's age, calculated from the date of birth to the date of thyroidectomy, and sex. Clinical presentations were recorded from the case notes, noting symptoms such as anterior neck swelling, difficulty swallowing, difficulty breathing, and voice change, along with the presence or absence of thyrotoxicosis. The size of the thyroid swelling was categorized according to the WHO grading system into Grades 0–2 [11]. Retrosternal extension and laterality of the disease were also documented. Postoperative histology reports were also assessed. The type of thyroid surgery performed and the length of surgery were abstracted from the theatre logs. The outcome variable was the occurrence of in-hospital post-operative complications, and the severity of the complications was classified according to the Clavien-Dindo classification [12]. These complications included hypocalcemia, laryngeal nerve injuries, and perioperative hemorrhage. However, definitive vocal cord paralysis and permanent hypocalcemia could not be assessed due to the limited in-hospital follow-up period.

Data analysis

Data was analyzed using the Statistical Package for Social Sciences (SPSS) version 27. Categorical variables were summarized into frequencies with proportions, while continuous variables were summarized into means with standard deviation or medians with the inter quartile range. We performed a univariable logistic regression analysis to assess which factors were associated with in-hospital post-thyroidectomy complications and to obtain crude odds ratios (cORs) with corresponding 95 % confidence intervals (95 % CI). We then performed a multivariable logistic regression analysis, including all factors with p-values <0.2 in the univariable analysis, to identify independent predictors of in-hospital outcomes after thyroidectomy.

Ethics approval

The study protocol was reviewed and approved by the Institutional Review Board of MUHAS IRB number (MUHAS-REC-07–2023–1822). The study adhered to the Helsinki Principles for conducting research involving human subjects. No direct patient identifiers were used during data analysis after de-identification. Direct patient identifiers used during data abstraction were de-identified during data transfer into SPSS to maintain patient anonymity.

Results

Description of study participants

In Table 1 below, we present a summary of the clinical demography of the study participants. Between January 2018 and December 2023, 405 patients underwent thyroidectomy at the study institution. The median age at surgery was 48 years (IQR 18–83), with nearly half of the cohort (48.4 %) aged 40–59 years. Women constituted the vast majority of cases (89.6 %), yielding a female-to-male ratio of 8.6:1. Anterior neck swelling was the predominant clinical presentation, reported in 98.5 % of patients. Most individuals presented with Grade 2 goiters, and more than half (58.5 %) exhibited bilateral thyroid lobe involvement. Retrosternal extension was observed in 49 patients (24.0 %). Comorbid conditions were documented in 108 patients (26.7 %), with hypertension (79.6 %) and type II diabetes mellitus (19.4 %) as the most prevalent. The disease course was generally longstanding, with a median symptom duration of 36 months (IQR 12–72). Beyond anterior neck swelling, the most frequent accompanying symptoms were thyrotoxic features (22.2 %), voice changes (11.9 %), and dysphagia (11.9 %). The majority of patients (77.0 %) had benign thyroid disease, and the median operating time was 120 min (IQR 80–150).

Table 1.

Clinical and demographic characteristics of patients who underwent thyroidectomies at MNH between 2018 and 2023.

Parameter	Frequency (Percent)	Median (IQR)
Age N=405		48 (IQR 18–83)
18 - 39	121 (29.9)
40 - 59	196 (48.4)
≥60	88 (21.7)
Sex N=405
Female	363 (89.6)
Male	42 (10.4)
Presentation N=405
Anterior neck swelling	399 (98.5)
Difficulty in swallowing	48 (11.9)
Difficulty in breathing	41 (10.1)
Voice change	48 (11.9)
Thyrotoxicosis	90 (22.2)
Size of the swelling N=405
Grade 0	7 (3.2)
Grade 1	9 (3.2)
Grade 2	389 (96)
Retrosternal extension N=405
Yes	49 (24)
No	155 (76)
Laterality N=405
Unilateral	168 (41.4)
Bilateral	237 (58.5)
Comorbidities N=108
DM	21 (19.4)
HTN	86 (79.6)
HIV	9 (8.3)
Others	18 (16.7)
Histology N=405
Benign	312 (77.0)
Malignant	93 (23.0)
Surgery Duration N=405		120 (80 – 150)

Types of thyroidectomies performed at MNH between 2018 and 2023

Fig. 1 summarizes the different types of thyroid surgeries performed. Lobectomy was the most commonly performed procedure, accounting for 27.2 % (n = 110) of cases, followed closely by total thyroidectomy in 24.0 % (n = 97). Other less frequently performed procedures included the Dunhill procedure, isthmectomy, and nodulectomy, collectively representing 4.0 % of all surgeries. Only 3.5 % of cases had a documented lymph node dissection performed during thyroidectomy.

Fig. 1.

Types of thyroidectomies done at MNH between 2018 and 2023.

Post-operative complications amongst patients who underwent thyroidectomies at MNH between 2018 and 2023

Postoperative complications occurred in 15.3 % of patients, as seen in Fig. 2A. As illustrated in Fig. 2B, the severity of complications was categorized using the Clavien-Dindo classification system. The majority of complications were classified as Grade I (56.5 %). Grade II and III complications accounted for 20.3 % and 11.6 % of cases, respectively. More severe complications were less frequent, with Grade IV complications observed in 5.8 % of patients and Grade V in 2.9 %.

Fig. 2.

Pie-charts showing the proportion (2A) and severity (2B) of post-thyroidectomy in-hospital complications amongst patients who underwent thyroidectomies.

Distribution of the post-thyroidectomy in-hospital complications according to severity

Fig. 3 shows the distribution of specific postoperative complications based on the Clavien-Dindo classification. Among Class I complications, hypocalcemia was the most common (4.0 %), followed by superior laryngeal nerve injury (3.0 %), postoperative hemorrhage (1.5 %), and recurrent laryngeal nerve injury (1.2 %). In Class II, postoperative bleeding remained notable at 1.5 %, with hypocalcemia and recurrent laryngeal nerve injury each contributing 1.0 %. Class III complications were mainly due to recurrent laryngeal nerve injury (1.0 %) and postoperative bleeding (0.7 %). Class IV events were solely attributed to severe postoperative bleeding (1.0 %), while Class V events (mortality) occurred in 0.5 % of cases. The deaths were associated with recurrent laryngeal nerve injury leading to prolonged intubation, with pneumonia documented as the final cause of death.

Fig. 3.

Distribution of the in-hospital post-thyroidectomy complications according to severity.

Univariable analysis of factors associated with the occurrence of post-thyroidectomy in-hospital complications

Table 2 below shows the univariable analysis of how sociodemographic and clinical characteristics relate to post-operative complications. In terms of age, older age was significantly associated with an increased risk of postoperative complications, whereby 25 % of patients aged 60 years or older experienced a complication, compared with 14.3 % among those aged 40–59 years and 9.9 % among those under 40 (p = 0.01). Duration of symptoms also demonstrated a stepwise increase in complication rates, rising from 10.0 % among patients with symptoms lasting <1 year to 13.8 % among those with symptoms lasting 1–10 years, and 15.3 % among those with symptoms lasting >10 years (p < 0.001). The laterality of thyroid involvement was also notable, with patients having bilateral disease experiencing complications in 19.8 % of cases, compared to 8.9 % in those with unilateral involvement (p = 0.003). Surgical extent was among the most significant variables: total thyroidectomy was associated with the highest complication rate (27.8 %), while lobectomy had the lowest (8.2 %) (p = 0.001). Although patients undergoing procedures lasting 120 min or longer experienced a higher complication rate (18.3 %) than those with shorter operations (11.8 %), this difference did not reach statistical significance (p = 0.073). Other variables, including patient sex, comorbidities, histologic classification (benign vs. malignant), and preoperative anemia, were not significantly associated with postoperative complications.

Table 2.

Univariate analysis of factors associated with the occurrence of post-thyroidectomy in-hospital complications.

Variable	Post-Thyroidectomy Complication		P Value
	“Yes” n ( %) 62(15.3)	“No” n ( %) 343 (84.7)
Age			0.01
60 +	22(25)	66(75)
40 – 59	28(14.3)	168(85.7)
18 – 39	12(9.9)	109(90.1)
Comorbidity			0.352
Yes	20(18)	91(82)
No	42(14.3)	252(85.7)
Duration of Symptoms			<0.001
> 10 years	62(15.3)	33(64.7)
1 – 10 years	31(13.8)	193(86.2)
< 1 year	13(10)	117(90)
Size of Goiter			0.17
Grade 0	0 (0.0)	7 (100.0)
Grade 1	3(33.3)	6(66.7)
Grade 2	59(15.2)	330(84.8)
Goiter extent			0.003
Unilateral	15(8.9)	153(91.1)
Bilateral	47(19.80)	190(80.2)
Histology			0.288
Malignant	11(11.8)	82(88.2)
Benign	51(16.3)	261(83.7)
Extent of Thyroidectomy			0.001
Lobectomy	9(8.2)	101(91.8)
Near-total	10(21.3)	37(78.7)
Total	27(27.8)	70(72.2)
Subtotal	2(5.1)	37(94.9)
Hemithyroidectomy	9(13)	60(87)
Completion	2(7.4)	25(92.6)
Others	3(18.8)	13(81.2)
Sex			0.497
Female	54(14.9)	309(85.1)
Male	8(19)	34(81)
Pre-Op Anemia			0.994
Yes	26(15.3)	144(84.7)
No	36(15.3)	199(84.7)
Surgery duration			0.073
<120mins	22(11.8)	164(88.2)
≥120mins	40(18.3)	179(81.7)

Multivariate analysis of factors associated with the occurrence of post-thyroidectomy in-hospital complications

Table 3 shows the multivariable analysis of predictors of post-thyroidectomy complications. Symptom chronicity and the extent of surgery emerged as independent predictors of postoperative complications. Patients with symptoms for >10 years had a four-fold increased likelihood of experiencing complications (aOR 4.12; 95 % CI 1.60–10.58; p = 0.003). Similarly, total thyroidectomy was associated with a three-fold increase in complication risk compared to lobectomy (aOR 3.12; 95 % CI 1.08–9.02; p = 0.035). Notably, age ≥60 years, which was significant in univariate analysis, lost significance after adjustment for other variables (aOR 1.59; 95 % CI 0.63–4.03; p = 0.33).

Table 3.

Multivariate analysis on factors associated with the development of post-thyroidectomy in-hospital complications.

Variable	cOR (95 % CI)	P-value	aOR (95 % CI)	P-value
Age (years)
18 – 39	Ref
40 – 59	1.5(0.7 – 3.10)	0.258	1.18(0.53 - 2.62)	0.68
Above 60	3.03(1.4 – 6.5)	0.005	1.59(0.63 – 4.03)	0.33
Duration of symptoms
< 1 year	Ref
1 – 10 years	1.4(0.73 – 2.87)	0.293	1.9(0.86 - 4.215)	0.111
> 10 years	4.9((2.18 – 11.04)	<0.001	4.115(1.60 - 10.58)	0.003
Laterality
Unilateral	Ref
Bilateral	2.5(1.36 – 4.7)	0.003	1.57(0.69 - 3.57)	0.286
Extent of thyroidectomy
Lobectomy	Ref
Near-total	3.03(1.14 – 8.05)	0.026	2.03(0.59 - 6.98)	0.263
Total	4.33(1.9 – 9.76)	<0.001	3.12(1.08 - 9.02)	0.035
Subtotal	0.6(0.13 – 2.94)	0.535	0.72(0.13 - 3.92)	0.702
Hemithyroidectomy	1.68(0.63 – 4.47)	0.296	1.57(0.5 - 4.9)	0.441
Completion	0.9(0.18 - 4.42)	0.894	0.88(0.16 - 4.94)	0.886
Others	2.59 (0.62 – 10.8)	0.192	2.98(0.59 - 15.15)	0.187

Discussion

To the best of our knowledge, this is the first study from Tanzania and among the most extensive series reporting thyroidectomy outcomes in East Africa. Drawn from a high-volume quaternary referral center over five years, the cohort offers an analysis of surgical practice in a pragmatic, resource-limited setting. These findings can inform surgical care pathways, facilitate benchmarking of complication rates, and guide quality improvement initiatives across the region, particularly in institutions operating under similar settings.

The mean age of patients undergoing thyroidectomy in our cohort was 48 years, with nearly half of the patients falling in the 40–59 age group. This is comparable to studies from similar sub-Saharan contexts, such as Nigeria, where the middle-aged age group is also the most commonly affected [13]. However, in high-income countries such as the United States, thyroid disease tends to be more prevalent among older adults (aged 60 and above), suggesting regional variability in age distribution that could reflect differences in predisposing factors and health-seeking behavior [14]. Conversely, the gender distribution showed a high female predominance, mirroring global patterns, where thyroid disorders disproportionately affect women. The predominance of females amongst patients with thyroid diseases has been documented in other regional and international studies, with female to male ratios ranging widely from 3:1 to 16:1. This difference has been attributed to a potential underlying hormonal or genetic influence in the development of thyroid diseases [8,[15], [16], [17]].

In terms of clinical characteristics, the finding that 98.5 % of patients presented with anterior neck swelling is expected and aligns with global literature, where thyroid masses are the primary clinical presentation of thyroid conditions [[18], [19]]. We also observed that one out of four had a retrosternal extension, which is consistent with data from international literature, where retrosternal extensions complicate the surgical approach and are observed in about 3–20 % of cases [[20], [21], [22], [23]]. Approximately one in four patients had at least one comorbidity, most commonly hypertension and diabetes mellitus, a profile typical of adult populations in sub-Saharan Africa [24]. This coexistence of thyroid and metabolic diseases is well-documented and may reflect both shared risk factors and the broader burden of non-communicable diseases in the region [25,26].

Benign conditions accounted for the majority of thyroidectomies in our cohort, accounting for almost 8 out of 10 cases. These findings are in line with reports from other LMICs, where benign thyroid pathology predominates [27]. Similarly, studies from Yemen, Turkey, and the U.S. also demonstrate that over 80 % of patients undergoing thyroidectomy have benign disease, supporting the notion that the natural history of thyroid enlargement in iodine-deficient or endemic regions tends toward benign hyperplasia [[28], [29], [30]]. Additionally, the median duration of symptoms before surgery was three years, highlighting persistent delays in diagnosis and treatment. Chronic symptom duration is commonly reported in LMICs and is often due to a combination of limited access to specialized care, low public awareness, and under-referral from primary healthcare levels [15,31,32].

The reported overall complication rate following thyroidectomy in this study was 15.3 %, which aligns with findings from various regional and international studies but contrasts with some others. For instance, a study from Italy reported that 17.4 % of thyroidectomy patients presented with postoperative complications [33]. Conversely, a study in the United States showed a significantly lower overall complication rate of only 3.28 % [34]. The majority of complications observed in our study were classified as Clavien-Dindo class I and II, indicating that most were mild and self-limiting. Among class I complications, hypocalcemia and superior laryngeal nerve injury were most prevalent. Hypocalcemia, in particular, is recognized as the most common post-thyroidectomy complication, especially following total thyroidectomy, where inadvertent devascularization or removal of the parathyroid glands is a known risk [33,[35], [36], [37]]. Postoperative hemorrhage was reported in nearly 4.7 %, which is within the reported global incidence range of 0–6.5 % [[38], [39], [40]].

Analysis of predictors of post-operative complications revealed that age over 60 years was significantly associated with postoperative complications in univariate analysis, a finding that aligns with previous studies suggesting age-related physiological decline and comorbidity burden may elevate surgical risk. However, this association lost statistical significance in the multivariate model. In contrast, symptom duration beyond 10 years was independently associated with higher complication rates. This supports the hypothesis that chronic disease leads to anatomical distortion and surgical difficulty, increasing the likelihood of intra- and postoperative events [35].

The extent of thyroidectomy also emerged as a significant predictor of postoperative complications. Total thyroidectomy, compared to lobectomy and other limited resections, was associated with a threefold higher risk of complications. This observation is consistent with literature from both high- and low-resource settings, where more extensive procedures have been linked to higher rates of nerve injury, hypoparathyroidism, and bleeding [[16], [33], [35], [36], [37], [38]] Although bilateral disease appeared to have higher complication rates than unilateral cases, the association was not statistically significant in adjusted analysis. Additionally, variables such as comorbidities, histology (benign vs malignant), operative time, and retrosternal extension were not significantly associated with complications in multivariate analysis, despite prior studies suggesting they are relevant risk factors [[16], [17], [18], [33], [34]]. This discrepancy highlights the potential influence of institutional factors such as surgical techniques, patient selection, and postoperative care protocols, warranting further studies.

While this study provides an evaluation of thyroid surgery outcomes in a sub-Saharan African context, it has some limitations inherent to its retrospective, single-center design. Potential sources of bias include incomplete documentation and limited control over confounding variables, particularly those related to surgical technique, intraoperative findings, and surgeon expertise, which were not accounted for in the study. The absence of long-term follow-up also precludes the evaluation of permanent complications, such as permanent hypoparathyroidism or recurrent laryngeal nerve palsy. Future research should aim to address these limitations by using prospective, multicenter cohort designs that incorporate standardized surgical protocols, stratification by surgeon experience, and long-term outcome tracking.

CRediT authorship contribution statement

Albert Lazaro: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Mungeni Misidai: Writing – review & editing. Daniel Kitua: Writing – review & editing, Methodology, Formal analysis. Abdulrahaman Amin: Writing – review & editing, Supervision. Charles Komba: Writing – review & editing, Supervision. Ally Mwanga: Writing – review & editing, Supervision. Larry Akoko: Writing – review & editing, Supervision. Nashivai Kivuyo: Writing – review & editing, Supervision, Methodology, Formal analysis, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.