A rare case of Haemophilus parainfluenzae septic knee following prolotherapy injections

Abstract

Prolotherapy injections are becoming increasingly popular as a non-surgical treatment option for many chronic musculoskeletal conditions. Proposed benefits include reduced pain, reduced joint laxity and increased tendon strength. While a number of studies report that prolotherapy reduces pain and increases function for many conditions, the academic evidence remains extremely weak. Here, we discuss a case of a complex intra-articular knee infection in a young, previously healthy, female following prolotherapy injections for management of a partial-thickness anterior cruciate ligament (ACL) tear. To the best of our knowledge, this is the first report of its kind describing a potential complication of intra-articular prolotherapy injections.

1. Background

Musculoskeletal pain has a large impact on quality-of-life. Whether originating from trauma, overuse or degenerative processes, effective treatment options are often limited, pushing health providers to come up with improved methods to provide relief to their patients. Prolotherapy was introduced as an injection therapy nearly a century ago (1937) and is becoming increasingly popular as a non-surgical treatment option for many musculoskeletal conditions.^1,3 Prolotherapy is performed by injecting an irritant, most commonly hyperosmolar dextrose or sodium morrhuate, into the peritendinous region or the joint space. However, the mechanism of action of prolotherapy is still not clear. The main hypothesis is that the intentional irritation promotes ligamentous shortening through an inflammatory response, causing an increase in growth factors and collagen production.^4,5

Several studies have examined the efficacy of prolotherapy in reducing pain and improving overall joint function in a variety of conditions, including unstable knees due to anterior cruciate ligament (ACL) laxity, knee osteoarthritis, Achilles tendinopathy, etc.^1,3 While a number of studies report that prolotherapy reduces pain and increases function, the academic evidence continues to be extremely weak. A large part of these studies are case reports or small case series containing no comparison groups. With very few randomized controlled trials or prospective studies, the ones that have been performed mostly have small sample sizes and poor randomization and/or poor blinding. Attempts to summarize and evaluate the literature have been made through systematic reviews, however these have all reached conclusions that more high-quality evidence is needed.^{5, 6, 7, 8, 9}

Despite insufficient evidence supporting prolotherapy as a clinically effective treatment option, it is becoming widely used across North America. Most likely because it is affordable, the injectable solution is easily obtained, and adverse events are rare.^{5, 6, 7, 8, 9} A literature search of the PubMed and MEDLINE databases (from inception to July 26th, 2018), did not identify any previous reports of infection or serious adverse events following intra-articular or peri-tendinous prolotherapy injections. The most common adverse events reported were transient bruising, local pain, and discomfort to the area following injection. So, for the patient, it may seem that the potential rewards far outweigh the minimal risks.

Here we report a serious joint infection with Hemophilus parainfluenzae in a young competitive female hockey player, following prolotherapy injections used to treat a partial-thickness ACL injury. H. parainfluenza is a gram-negative bacteria most commonly found in the oral-pharyngeal region. As far as the authors of this paper are aware, this is the first serious, adverse event with long-term consequences following intra-articular prolotherapy injections to be reported. This emphasizes the importance for an adequate assessment of the safety and efficacy of prolotherapy and questions whether its widespread use is warranted.

The patient was informed that data concerning the case would be submitted for publication, and she provided consent.

2. Case presentation

2.1. Emergency management

An 18-year old elite female athlete developed sudden onset of malaise and acute knee effusion. This was preceded by prolotherapy injections for left knee pain. The patient was told that these injections would help her symptoms, which were attributed to a partial-thickness ACL injury diagnosed by magnetic resonance imaging nine months prior. The patient played highly competitive hockey and was in the process of determining her choice of university hockey scholarship. Her family stated that the practitioner had explained that the injections would help impart healing factors in the knee that could settle the patient's pain. Shortly following her last injection, the patient experienced worsening generalized knee pain, and within 48-h, developed a significant knee effusion and was unable to bear weight, at which point she presented to our Emergency Department.

Her past medical history was only significant for a partial-thickness ACL tear. She was not currently taking any medications, including prescription or over-the-counter medications, and was a non-smoker. She had no family history of congenital immunodeficiency disorders nor had she ever manifested any signs or symptoms of frequent, chronic, or opportunistic infections during infancy and childhood.

She presented with severe pain in the affected joint, exacerbated with movement. On physical examination, a knee effusion was apparent and was painful to touch. She was very apprehensive to passive and active flexion and extension of her knee due to exacerbated pain. She did not complain of any fevers or chills. Her vital signs upon presentation were: temperature of 37.1 °C, a pulse of 103 beats per minute, respiration rate of 18/minute, blood pressure of 118/77 mmHg, and O₂ saturation of 92% on room air. Routine laboratory workup revealed a normal white blood cell count (7.3 10⁹/L), but elevated inflammatory markers (C-reactive protein, 181.5 mg/L; Erythrocyte Sedimentation rate, 77 mm/hr) (Table 1). A knee arthrocentesis was performed; fluid cell count and microbiology results can be found in Table 2. Anaerobic and fungal cultures were also performed. Due to our high suspicion for an intra-articular infection following the prolotherapy injection, consent was obtained to undergo arthroscopic irrigation and debridement (I & D) and a 4-compartment synovectomy.

Table 1.

Baseline bloodwork upon presentation to the emergency.

Test	Results	Ref. Range	Abnormality
C-Reactive Protein	181.5 mg/L	0.0–8.0 (mg/L)	High
Sedimentation Rate	77 mm/hr	0-20 (mm/hr)	High
Creatinine, serum	62 μmol/L	35-100 (umol/L)
Anion Gap	11 mmol/L	4-16 (mmol/L)
Lactate, blood gas	1.8 mmol/L	≤2.0 (mmol/L)
Base excess	−2 mmol/L	-5-1 (mmol/L)
White Blood Cell, serum	7.3 109/L	4-11 (109/L)

Table 2.

Fluid cell count and microbiology, synovial fluid aspirate.

Test	Results	Ref. Range	Abnormality
Results from first I & D
Synovial fluid source	Synovial, knee
Synovial fluid clarity	Cloudy
Synovial fluid color	Yellow
Synovial fluid WBC	17.9 109/L	0.0–0.2 (109/L)	High
Synovial fluid RBC	<0.01 1012/L	(1012/L)
Fluid culture report	Haemophilus parainfluenzae		Abnormal
Suseptability	Ampicillin, Ceftriaxone, Cefuroxime, Ciprofloxacin
Results form second I & D
Synovial fluid source	Synovial, Knee
Synovial fluid clarity	Cloudy
Synovial fluid color	Slightly bloody		Abnormal
Synovial fluid WBC	36.3 109/L	0.0–0.2 (109/L)	High
Synovial fluid RBC	<0.03 1012/L	(1012/L)
Fluid culture report	Haemophilus parainfluenzae		Abnormal
Suseptability	Ampicillin, Ceftriaxone, Cefuroxime, Ciprofloxacin

2.2. Surgical management

The intra-articular assessment was consistent with a purulent bacterial infection with inflamed synovium. Also noted was several punctate bleeding points on the ACL which were consistent with several puncture wounds from the prolotherapy injections (see Fig. 1). Both menisci were felt to be stable without any appreciable tears. Intraoperative synovial fluid was sent to microbiology. She was later assessed by the Infectious Disease team and started on scheduled Cefazolin 2 g intravenously, every 8 h. She was not formally assessed for any immunodeficiency, given her clinical profile. The final bacterial cultures (reported 72 hours after collection) proved to be Haemophilius parainfluenzae, and therefore, given the susceptibility results, Cefazolin was replaced with Ciprofloxacin 750 mg orally, given twice daily (Table 2).

Fig. 1.

Intraoperative arthroscopic pictures taken during the second irrigation and debridement of the patient's knee. An irrigation and debridement and 4-compartment synovectomy was completed. A and B) Hemorrhagic staining of the ACL from the previous prolotherapy injections. C) Friable articular cartilage of the femoral condyle which was felt to be due to the ongoing intra-articular infection. D) Lateral compartment had hemorrhagic material with an intact meniscus.

Due to persistent symptomatology for 72 hours following the initial surgery, we elected to proceed with a second arthroscopic I & D and further synovectomy in order to lower the infectious material count. Ongoing infection was easily confirmed by the presence of infectious and hemorrhagic material evacuated from the joint while establishing the portals. An intraoperative sample of synovial fluid was sent to Microbiology (Table 2). Nine litres of saline, with Bacitracin in the final bag was used. A thorough four-compartment synovectomy was completed. The medial and lateral gutter, intercondylar notch, and bilateral compartments all had hemorrhagic material, as well as friable articular cartilage that was felt to be due to the ongoing infection. She was discharged on post-operative day two, where she was to complete a 4-week course of Ciprofloxacin.

Upon her first clinical follow-up (post-operative day 10), there was no evidence of clinical improvement - possibly due to the atypical pathogen isolated. Our concern was the possibility of a polymicrobial infection, which would have been far more challenging to treat. In consultation with the Infectious Disease specialist, we decided to proceed with a third arthroscopic I & D. This time, several intraoperative tissue samples were collected along with synovial fluid and sent to the Microbiology laboratory for further cultures and sensitivities. After 4 days of incubation, these returned with no growth of any organisms. Postoperatively, the patient was started on Ceftriaxone 2 g intravenously, daily for six-weeks as definitive antibiotic management.

During the months that followed, the patient slowly regained her range of motion and most of her strength. At her two-year follow-up, the patient admits to experiencing ongoing mild symptomatology - specifically patellofemoral pain. She now works as a hockey coach, as she has been unable to return to competitive hockey.

3. Discussion

To the best of our knowledge, this is the first case reporting a serious life-changing infection following prolotherapy treatment. Here, an 18-year old female with no significant past medical history developed a severe knee infection due to H. parainfluenzae following prolotherapy injections to her ACL. The incidence of infection following any type of intra-articular joint injection is rare; being estimated somewhere between 1:3000 to 1:50,000 cases.¹⁰ Similarly, H. parainfluenzae is an extremely rare cause for septic arthritis in adults, with an incidence of less than 1% of cases being caused by any type of Haemophylus sp, and only 12 published cases reporting a positive culture for H. parainfluenzae.^{11, 12, 13} Furthermore, H. parainfluenzae in adults usually presents with underlying comorbidities in the context of an immunocompromised condition, such as chronic alcoholism, cardiopulmonary disease, HIV infection, or malignancy.¹³

We hypothesize that, in the presented case, the prolotherapy injections greatly increased the risk of infection, either by direct introduction of the bacteria through improper aseptic technique, and/or by creating an environment that promoted bacterial growth introduced through the bloodstream, as suggested by Borenstein and Simon (1986).¹¹ The most common irritant solution used for prolotherapy is hypertonic dextrose. This sugar-based irritant, in combination with the small hematomas induced by the trauma of the injections and the decreased blood flow naturally present in ligamentous tissue, would provide a prime medium for bacteria to flourish. This is analogous to how suspected infections are tested for and are grown in laboratory settings – via incubation in a petri dish, in the presence of blood and nutrient sugar. The combined micro-trauma and introduction of dextrose may therefore substantially increase the risk of infection.

Given the high annual frequency of ACL injuries in the United States (∼81/100,000 people/year), any effective non-operative management of ACL symptoms is of significant interest to the general orthopaedic community.¹⁴ Unfortunately, current literature addressing the safety and efficacy of intra-articular prolotherapy injection is scarce, and the absence of high-quality methodology among prior published studies prevents the generation of quality evidence-based clinical recommendations.^2,6,7,9,15

We therefore highly recommend that other evidence-based treatments for musculoskeletal disorders - including physiotherapy, lifestyle interventions and/or surgical management - be exhausted prior to consideration of prolotherapy treatments. Further studies on prolotherapy should also carefully consider the potential for adverse events such as infection.

4. Conclusions

At this time, there is insufficient high-quality literature to make evidence-based recommendations regarding the use prolotherapy for the management of partial-thickness ACL injuries. Given the lack of demonstrated efficacy and the risk of infectious complications, treatment with prolotherapy interventions are not indicated at this time. This case report highlights the risk of the potentially catastrophic infectious complications precipitated by prolotherapy treatment. Careful risk consideration and optimal technique should be undertaken prior to any intra-articular treatment.

Learning points

• 1There is presently a lack of high-quality evidence to make recommendations regarding the use prolotherapy for the management of partial-thickness ACL injuries.
• 2Infection is a potential complication of any intra-articular injection, including prolotherapy injections. Despite being uncommon, caution is warranted and meticulous technique is strongly encouraged.
• 3Careful risk consideration must be discussed with patients.

Copyright statement

I, Stephen French, The Corresponding Author, has the right to assign on behalf of all authors and does assign on behalf of all authors, a full assignment of all intellectual property rights for all content within the submitted case report (other than as agreed with the BMJ Publishing Group Ltd) (“BMJ”)) in any media known now or created in the future, and permits this case report (if accepted) to be published on BMJ Case Reports and to be fully exploited within the remit of the assignment as set out in the assignment which has been read. http://casereports.bmj.com/site/misc/copyright.pdf.

Disclaimer

Despite numerous attempts, the authors were unsuccessful in gaining access to any technique or procedural notes by the practitioner who completed the prolotherapy injections.

Appendix A. Supplementary data

The following is the Supplementary data to this article: