Postoperative serum C-reactive protein dynamics after pharyngolaryngectomy with jejunal free-flap reconstruction

Abstract

Introduction

Pharyngolaryngectomy with jejunal free-flap (JFF) reconstruction can be offered for locally advanced hypopharyngeal cancer. However, the procedure carries significant morbidity. Postoperative serial serum C-reactive protein (CRP) has been shown to be a marker predicting postoperative complications, and the aim of this study was to describe the dynamics and value of CRP in this patient group.

Methods

Retrospective analysis of pharyngolaryngectomies with JFF reconstruction was performed in our institution. Daily postoperative CRP values were analysed within the first 14 days, as were complications.

Results

Twenty-one cases were included. Total morbidity was 57.1% including 14.3% (temporary) anastomotic leaks and 14.3% flap failures. Patients in the normal group showed peak CRP levels around postoperative day 2 (2.2). Increased CRP levels on or after day 4 were associated with complications (p<0.01) with a sensitivity of 83.3% and specificity of 77.8%. In keeping with CRP kinetics from other surgical studies, peak CRP values on day 2 or 3 are expected, followed by a decline. Peaks in CRP on day 4 or later raise the suspicion of complications. CRP is not specific for any one complication but rather can help guide early appropriate clinical assessment and management.

Conclusions

The natural postoperative CRP response peaks around postoperative day 2 (2.2) and declines thereafter. Rising CRP levels after postoperative day 3 are suspicious of surgical complications (p<0.01) with positive and negative predictive values of 83.3% and 77.8%, respectively. Therefore, serial postoperative CRP can be used as an adjunct to monitor outcomes in this group.

Introduction

Hypopharyngeal squamous cell carcinomas usually present late in advanced stages with poor survival rates.¹ Pharyngolaryngectomy can be offered for locally advanced disease when circumferential pharyngeal resection is required and immediate defect reconstruction with jejunal free flap (JFF) is an established robust method first described in 1959 by Seidenberg.² A recent meta-analysis supported JFF reconstruction showing significantly lower fistula and stricture rates compared with anterolateral thigh or radial forearm flaps.³ However, pharyngolaryngectomy with simultaneous laparotomy carries relevant perioperative morbidity.⁴

A special consideration is that the JFF is buried in the neck and current means of postoperative flap monitoring have limitations.⁵ JFF failure is a much-feared complication, potentially leading to delays in adjuvant treatment, carotid artery erosion/bleed or even death, and sufficient means of monitoring and detecting postoperative complications in this group are desirable. Serum C-reactive protein (CRP) is produced by the body in response to infection, inflammation, malignancy and trauma, and is widely used as a clinical marker of infection. To the best of the authors' knowledge, few data are available describing the usefulness of serial CRP levels in predicting early postoperative complications for head and neck patients.⁶ The aim of this study was to describe the dynamics and value of postoperative serial CRP to predict early head and neck complications in patients undergoing pharyngolaryngectomy with JFF reconstruction.

Methods

Retrospective analysis of pharyngolaryngectomies with JFF reconstruction was performed in our institution between 2011 and 2020. Patient records including operating records were reviewed. Twenty-one cases were included. Patients with elevated CRP levels due to nonsurgical-related systemic complications, such as hospital-acquired pneumonia, or relevant abdominal complications within the first 14 postoperative days were excluded. Daily postoperative CRP values were analysed within the first 14 days, as were JFF outcomes and early head and neck complications. Patients were separated into a normal group and a complicated group. CRP dynamics were classified into falling CRP on day 4 (post-surgery) and later, and rising CRP on day 4 or later over at least two consecutive days. Differences between the subgroups were analysed using the chi-square test. In addition, length of postoperative hospitalisation (in days) for the normal and complicated groups was analysed.

Results

Between 2011 and 2020, 21 cases were included. The patients’ characteristics are given in Table 1.

Table 1 .

Characteristics of patients who underwent pharyngolaryngectomy with jejunal free-flap reconstruction for hypopharyngeal squamous cell carcinoma

	Absolute number (n=21)	Relative %
Gender
Male	17	80.9
Female	4	19.1
Age, mean (years)	66.1 (range 40–80)
Smoking history
Yes	19	90.5
No	2	9.5
Histology
Squamous cell carcinoma	21	100
Previous treatment
Primary surgery	19	90.5
Salvage surgery	2	9.5
T-stage
T1	0	0
T2	1	4.8
T3	6	28.6
T4a	14	66.7
N-stage
N0	6	28.6
N+	15	71.4
Neck dissection
No	2	9.5
Unilateral	10	47.6
Bilateral	9	42.9

The specifics of all 21 patients are listed in Table 2. Total morbidity, as given in Table 3, was 57.1% (12/21). Three patients who demonstrated an early (temporary) anastomotic leak on contrast swallow healed spontaneously without further intervention or fistula. Of these three patients, only one showed raised CRP after day 4. One patient with a chylothorax required chest drainage. Of three patients in whom the JFF was lost, despite efforts at secondary reconstruction: one died 26 days after initial surgery; one had a persistent pharyngeal defect and developed a common carotid artery haemorrhage after 4 months, which was coiled by the interventional radiologist leading to an ischaemic stroke; and one was discharged after 172 days following multiple reconstructive surgeries, finally managing an oral diet. Unfortunately, this last patient missed the 6-week window for adjuvant radiotherapy through the course of recovery.

Table 2 .

Specifics of patients who underwent pharyngolaryngectomy with jejunal free-flap reconstruction for hypopharyngeal squamous cell carcinoma

No.	TNM	Prior CRT	Previous head and neck surgery	Operation	Operation time (min)	Complication	Hospital days
1	T4aN1M0	No	No	PL+L ND+JFF	554	No	18
2	T4aN2aM0	RT 26 years ago	No	PL+B/L ND+JFF	636	Temporary anastomotic leak	19
3	T3N0M0	No	No	PL+B/L ND+JFF	570	Temporary anastomotic leak	31
4	T4aN2cM0	RT 30 years ago	No	PL+B/L ND+JFF	510	Neck wound collection/infected seroma	37
5	rT2N0M0	CRT	Yes	PL+JEJ	525	JFF failure	81
6	T4aN3bM0	No	No	PL+B/L ND+JFF	662	Temporary anastomotic leak	35
7	T4aN2aM0	No	No	PL+B/L ND+JFF	817	JFF failure	172
8	T3N2bM0	CRT	Yes	PL+R ND+JFF	644	No	28
9	T3N2aM0	No	No	PL+R ND+JFF	625	No	38
10	T4aN1M0	No	No	PL+B/L ND+JFF	667	No	13
11	T4aN2cM0	No	No	PL+B/L ND+JFF	696	Chyle leak with chylothorax	24
12	T4aN2bM0	No	No	PL+L ND+JFF	560	No	19
13	T3N0M0	No	No	PL+B/L ND+JFF	639	No	25
14	T4aN2bM0	No	No	PL+R ND+JFF	605	JFF failure	died on day 36
15	T3N2cM0	No	No	PL+B/L ND+JFF	619	No	27
16	T4aN2bM0	No	No	PL+R ND+JFF	746	Wound abscess/infected seroma	16
17	T4aN0M0	No	No	PL+R ND+JFF	616	Neck wound cellulitis	21
18	T4aN2bM0	No	No	PL+L ND+JFF	590	No	24
19	rT3N0M0	CRT	No	PL+JFF	580	Neck wound collection/abscess	58
20	T4aN0M0	No	No	PL+L ND+JFF	545	No	24
21	T4aN2bM0	No	No	PL+R ND+JFF	693	Infected seroma neck	17
					Mean 624 min (range 510–817)		Mean 36.3 days (range 13–172)

B/L = bilateral; CRT = chemoradiotherapy; JFF = jejunal free flap; L = left; ND = neck dissection; PL = pharyngolaryngectomy; R = right; RT = radiotherapy.

Patients 5 and 19 were salvage cases. Patient 2 had undergone radiotherapy 26 years ago for lymphoma and patient 4 had undergone RT 30 years ago for breast cancer. Patient 8 had previously undergone resection of an oral cavity cancer with adjuvant CRT 2 years earlier.

Table 3 .

Early (within 14 days) local head and neck complications after pharyngolaryngectomy with jejunal free-flap reconstruction in a collective of 21 patients

Complication	Absolute number	Relative %
Neck collection/ infected seroma	4	19.0
Temporary anastomotic leak	3	14.3
JFF failure	3	14.3
Neck cellulitis	1	4.8
Chylothorax	1	4.8
Total	12/21	57.1

JFF = jejunal free flap

Statistical analysis, as given in Table 4, revealed a significant association between increasing CRP on or after day 4 and early postoperative complications (p<0.01) when compared with decreasing CRP levels in the normal group. The sensitivity was 83.3% and the specificity was 77.8%. The positive predictive value was 83.3% and the negative predictive value 77.8%.

Table 4 .

Relationship between postoperative C-reactive protein values in normal and complicated patients within 14 days after pharyngolaryngectomy with jejunal free-flap reconstruction

	Falling CRP on and after day 4	Rising CRP on or after day 4	p-value
Normal group	7	2	<0.01
Complicated group	2	10
Total	9	12

CRP = serum C-reactive protein

Figure 1 shows averaged CRP dynamics in the normal and complicated groups. The average CRP peak for patients in the normal group was on day 2 (2.2; range 1–4). In the complicated group, the highest average CRP value was recorded on day 4; however, individual CRP peaks were observed between day 2 and day 14.^4,6 Furthermore, in patients in the normal group, average CRP returned to below baseline levels on day 0 (operation day), whereas in the complicated group, CRP remained elevated over the whole observation period (Figure 1).

Figure 1 .

Average postoperative C-reactive protein (CRP) dynamics in normal and complicated patient groups within 14 days after pharyngolaryngectomy with jejunal free-flap reconstruction

Figure 2 demonstrates the peak CRP values recorded for normal, complicated and flap failure groups. The other complications group (nine patients including neck collection/infected seroma, temporary anastomotic leak, neck cellulitis and chylothorax) had wide-ranging CRP peaks from 59 to 396, and mean peak CRP was 228.33, compared with 187.88 in the normal group (nine patients) and 328.67 in the JFF failure group (three patients). The postoperative CRP peak in JFF patients was 75% higher than in normal group patients.

Figure 2 .

C-Reactive protein (CRP) peak levels (range and mean) for normal, complicated and flap failure patients within 14 days after pharyngolaryngectomy with jejunal free-flap (JFF) reconstruction

The mean length of hospitalisation is shown in Figure 3. The length of hospitalisation for the complicated group was almost twofold (1.9) that for patients in the normal group.

Figure 3 .

Length of hospitalisation (days) for patients in the normal and complicated groups after pharyngolaryngectomy with jejunal free-flap reconstruction

Discussion

Pharyngolaryngectomy with immediate JFF reconstruction requires a multidisciplinary team of skilled surgeons for resection, harvest of jejunal graft and microvascular anastomosis. The procedure requires many hours of meticulous work and, unfortunately, overall reported morbidity is as high as 78%⁴ with a 14-day complication rate of 57.1% in our series. Temporary anastomotic leaks found on contrast swallow usually heal spontaneously, and neck collections are managed simply by aspiration or drainage.

By contrast, JFF failures create major problems due to salivary leakage into neck tissues with exposed large vessels (internal jugular veins, carotid arteries) and require further reconstruction. All three patients in our series with JFF failure suffered adverse outcomes. Postoperative JFF monitoring still has limitations despite various techniques described in the literature, namely implantable Doppler probes, exteriorisation of a jejunal segment, watch windows, microdialysis, microendoscopy, reflectance photoplethysmographic and, more recently, external colour duplex ultrasound to monitor vascularity.⁵ None of the above methods have been shown to provide strong evidence so far and, in practice, clinical judgement needs to be exercised.

The literature reports fistula rates of between 0 and 32% and JFF failure rates of less than 10%.⁷ Our series revealed a higher JFF failure rate, which we assume was due to excluded patients (elevated CRP due to systemic complications, eg pneumonia), one salvage case and a change in the microvascular team during the observation period.

Plasma CRP was first isolated from patients with a pneumococcal pneumonia infection in 1930 and since then CRP estimation has become common clinical practice as an early indicator of infection.⁸ CRP is produced by the liver in response to infection, inflammation, malignancy and trauma, with a relatively rapid response, and is more predictable in the early postoperative period compared with the erythrocyte sedimentation rate.⁹ Kallio et al described elevated CRP levels in patients with open surgery for tibial fractures with peak levels recorded at day 2 after surgery¹⁰ and a second postoperative CRP increase, or failure to decrease, has been shown to be a marker for infection after spinal surgery.⁹ Furthermore, after abdominal surgery, serum CRP on day 4 has been shown to reliably exclude postoperative infectious complications, with a negative predictive value of 84.3% in a 2015 meta-analysis.¹¹

Plasma CRP increases even before clinical infection is apparent, and elevated CRP is highly suggestive of acute infection or trauma.¹² Plasma CRP is known to reach its peak value after about 48 hours and decrease sharply if the cause is treated.¹³

Studies of normal CRP kinetics in spinal, knee and cardiac surgery describe peak CRP values on postoperative day 2 or 3, followed by an initial sharp decline and then a gradual decrease with normalisation in postoperative days 14–21.⁶ In the reconstruction of lower limb with local and free flaps, CRP peaks were found on day 2 and peaks after postoperative day 4 indicated complications.¹⁴

Song et al described the usefulness of CRP levels for predicting flap complications after performing microvascular head and neck reconstruction in 25 patients.⁶ The average peak CRP value for normal patients was recorded on day 2.9 and highest CRP levels after postoperative day 4 were highly indicative of free flap complications. Flaps in this study included anterolateral thigh flap, radial forearm flap, vastus lateralis muscle flap, tensor fascia lata flap and thoracodorsal artery perforator flap after resection of different head and neck primaries.⁶

Another study described postoperative serial CRP screening as a useful indicator of infectious complications after oral cancer resection with primary reconstruction, with a negative predictive value of 100%, which means normal CRP responses could rule out almost all early infectious complications and a rise in CRP after postoperative day 3 was considered an indicator of infection.¹⁵ This study included surgical site infections as well, as non-wound infections such as pneumonia or urine infections.

Our study revealed CRP kinetics similar to that in Song et al with natural postoperative CRP peaks around day 2 (2.2) declining thereafter in the normal group patients and increasing CRP levels on day 4 or thereafter in the complicated group patients. In our study, the CRP increase was not specific for free flap failure, but was also seen in other postoperative complications. Nevertheless, despite the limited sample size, a correlation between abnormal CRP dynamics and adverse outcomes was seen. One should not forget that CRP is not a specific marker for any one complication, but rather can help guide appropriate clinical assessment and management. It can be used as an early warning sign of postoperative infectious wound complications. Knowledge and awareness of the natural and pathological CRP response in postoperative patients might be valuable for early detection of surgical site complications to hopefully prevent adverse outcomes, especially in flap failure patients. Therefore, after excluding other clinical causes of postoperative abnormal CRP (eg pneumonia, seroma), flap failure should be excluded by appropriate means, for example, colour Doppler ultrasound⁵ or ultimately surgical exploration. In addition, higher CRP peak values might be more indicative of free-flap failure. By contrast, falling CRP levels on postoperative day 4 and after have value in excluding postoperative complications, except the mentioned temporary anastomotic leaks in our series, which seem to cause less of an inflammatory response. In this context it is also of note that CRP testing is relatively cheap (£1.70 per test in our institution) and almost universally quickly available to augment postoperative monitoring of head and neck patients.

Conclusions

Pharyngolaryngectomy with JFF reconstruction is a major operation with significant 14-day morbidity (57.1%). Owing to the surgical trauma, there is a relevant natural CRP response, which peaks around day 2 (2.2) and declines thereafter. Increasing CRP levels after day 3 are suspicious of early surgical complications (p<0.01) with positive and negative predictive values of 83.3% and 77.8%, respectively. Rising CRP levels on day 4 or after should therefore initiate a thorough clinical assessment to determine the cause and facilitate appropriate early intervention. Therefore, serial postoperative CRP can be used as a valuable but cheap adjunct in postoperative monitoring for patients undergoing this head and neck surgery with free flap reconstruction.